This paper investigates the collaborative unmanned aerial vehicle (UAV) sensing in multi-UAV networks. By equipping sensors, communication units and computation units on UAVs, they can sense the environment by executing surveillance tasks and computation tasks. UAVs execute surveillance tasks by mon-itoring the environment and computation tasks by processing the sensing information offloaded from ground terminal nodes. We formulate the collaborative multi-DAV sensing problem as a joint optimization problem of UAV deployment and task allocation. The goal is to maximize the sensing utility, which is defined as the weighted sum of utility for executing surveillance and computation tasks. Note that UAV deployment and task allocation are highly coupled, we divide the joint optimization problem into two sub-problems and then propose an iterative mechanism by optimizing two sub-problems iteratively. UAV deployment optimization problem is non-convex, and thus we utilize a differential evolution (DE) algorithm to solve it with high efficiency and simple implementation. Task allocation optimization problem is proven to be NP-hard, and thus we propose a particle swarm optimization method with greedy-based reconstruction (PSOGR) to solve this complex combinatorial optimization problem efficiently. Simulation results demonstrate the benefits of our proposed method compared with existing algorithms for different system parameters.