Track-before-detect (TBD) technology based on the probability hypothesis density (PHD) filter can effectively solve the problem of tracking dim varying number multitarget. The existing PHD-TBD algorithm has two shortcomings, lack of accuracy in the number of targets and long time delay in responding to the targets being detected. The paper studied the PHD-TBD method, deduced the accurate expression of the updated particle weight of the PHD-TBD algorithm, and achieved the precise estimate of the number of targets. Simultaneously, by using Bayesian theory, it deduced the probability density sampling function of new born particles based on measurement, which can quickly and effectively find the targets. In addition, the simulation results demonstrate that the proposed algorithm can effectively estimate the number of targets, detect the targets and accurately estimate their positions with a more rapid speed compared with the existing PHD-TBD algorithm.