Given a video sequence, moving objects usually contain important information. So, detecting moving objects becomes the most significant part of various applications. In computer vision, the detection of moving objects from a video sequence based on moving objects is crucial in many visionbased applications such as action recognition, traffic controlling, industrial inspection, and human behavior identification. There is much research that has been done for detecting moving objects by the stationary camera. But a moving camera brings new challenges to moving object detection. Recently several methods for background subtraction from moving cameras were proposed. The background is often obtained by dominant single or multiple planes with a complex BG/FG probabilistic model. Some of them use bottom-up cues to segment video frames into foreground and background regions.They may fail to detect an object when the clues are ambiguous in the video. It is often due to this lack of explicit models. This article will discuss possible solutions to resolve the ambiguity in the moving camera's background subtraction problem and introduce factors that influence the BS's efficiency.