Target tracking requires both the identification of potential targets and the estimation of target trajectories. The tracking process utilizes the obtained information, i.e. measurements from sensors or radars, to solve the identification and estimation issues based on certain assumptions and models related to the target state propagation, the target-originated measurement generation processes and the clutter in the surveillance environment. There are zero or more targets in the surveillance area. The number and the position of the targets are prior unknown to the tracker. A basic assumption of the target-originated measurement generation is that a target can generate one measurement at each scan time, described by the probability of detection P_D which is usually less than unity. However, this assumption is not always valid especially in the scenarios where an extended target is tracked, or the over-the-horizon radar (OTHR) is used to detect the target that is beyond the radar horizon. As a result, a target can give rise to more than one measurement at each sampling time. An extended target generates multiple detections due to the fact that there are multiple signal scattering points within its extension. However, the target in the OTHR system is still treated as a point mass, and the multiple detections from a target are caused by the multiple signal propagation paths. Therefore, the multiple detection generation mechanisms behind these two situations are completely different. Even in the multiple detection scenarios, each measurement has only one source---``a measurement is from a target of interest'' or ``a measurement is from clutter''.The content proposed in this thesis is based on joint integrated probabilistic data association (JIPDA) to solve the multiple detection problem caused by the extended target or the utilization of the OTHR system. The results are multiple detection linear multitarget integrated probabilistic data association (MD-LM-IPDA), multiple detection Markov chain joint integrated probabilistic data association (MD-MC-JIPDA), and multi-path linear multitarget integrated probabilistic data association (MP-LM-IPDA). These algorithms are divided into two groups: 1) MD-LM-IPDA and MD-MC-JIPDA belong to the multiple extended target tracking and 2) MP-LM-IPDA is of the OTHR target tracking family.Extended target tracking: The extended target tracking problem, which occurs due to a high-resolution sensor resolving multiple scattering points of a target, is considered. The multiple detection joint integrated probabilistic data association (MD-JIPDA) firstly utilizes the measurement partition method to generate measurement cells. Then, these measurement cells are considered in the joint data association events for assigning measurement cells to tracks. Thus, track-to-measurements associations can be realized such that each track can be associated with more than one measurements. In MD-JIPDA, joint data association events are used to assign measurement cells to tracks considering all measurement cells and tracks. The number of feasible joint events grows exponentially with the number of measurement cells and the number of tracks. As a result, MD-JIPDA is plagued by large increases in the computational load in multiple detection scenarios, especially when targets are closely spaced. Multiple detection linear multitarget integrated probabilistic data association (MD-LM-IPDA) is proposed to enhance computational efficiency by incorporating the modulated clutter measurement density that takes into account the influences from clutter as well as other targets to a measurement cell. Therefore, MD-LM-IPDA can entirely bypass utilizing joint data association events. The computational complexity of MD-LM-IPDA is linear with the number of measurement cells and the number of tracks since each track propagates separately with coupling among individual tracks evaluated in the modulated clutter measurement density. The multiple detection Markov chain joint integrated probabilistic data association (MD-MC-JIPDA) is also designed to reduce the computational load, in which the Markov chain process is used to generate data association sequences for each track. The sequences of individual tracks form the substitutes of the joint data association events. The Markov chain process can significantly reduce the computational cost due to the fact that only a few number of association sequences are generated by each track. OTHR target tracking: In target tracking environments using over-the-horizon radar (OTHR), one target may generate multiple detections through different signal propagation paths. Trackers need to jointly handle the uncertainties stemming from both the measurement origin and the measurement path. Traditional multitarget tracking algorithms suffer from high computational loads in such environments since they need to enumerate all possible joint measurement-to-track assignments with consideration of the measurement paths. To reduce the computational load, some algorithms employ approximations regarding the measurements and their corresponding paths. The multi-path linear multitarget integrated probabilistic data association (MP-LM-IPDA) is designed to efficiently track multitarget using OTHR. MP-LM-IPDA calculates the modulated clutter measurement density for each measurement cell with a possible path pattern instead of enumerating all possible joint assignments among measurements, paths and tracks. The path pattern assigns non-repeated paths for the measurements in a measurement cell. Then, the modulated clutter measurement density is evaluated for each path pattern combined measurement cell, which considers the possibility that the measurement cell if originated from the clutter as well as from other potential targets. By incorporating the modulated clutter measurement density, the single target tracking structure can be applied for multitarget tracking, which significantly reduces the computational load. MP-LM-IPDA could be treated as the practical application of MD-LM-IPDA to OTHR system.