In multi-target tracking scenarios of unmanned aerial vehicles (UAVs), the high computational complexity becomes a serious problem for the cardinalized probability hypothesis density filter (CPHD) because of the dense targets and intensive clutter. This paper proposes a novel multi-target tracking algorithm termed decoupled unbiased converted measurements adaptive gating cardinalized probability hypothesis density (DUCM-AG-CPHD) algorithm to solve this problem. Firstly, the proposed algorithm utilizes the decoupled unbiased converted method to obtain more information about the polar coordinate measurements. Meanwhile, it compensates for the converted error caused by coordinate transformation. Secondly, a novel adaptive gating strategy is designed to eliminate uncorrelated measurements, which can reduce the computational complexity. Finally, the Gaussian mixture implementation of the proposed algorithm is adopted to achieve efficient performance. The simulation results indicate that the proposed algorithm yields a considerable reduction in computational complexity and improves the estimated precision of both the number of targets and the states.