In order to solve the problem of particle impoverishment and low computational efficiency of a particle filter, an adaptive particle filter which can dynamic adjust the particle set size based on Kullback Leibler distance(KLD) is presented in this paper. Firstly, KLD sampling is used to dynamiclly adjust the particle set size in importance sampling. Then, the bat algorithm is used to optimaize the particle set purposefully, and it interacts with KLD sampling in iterative updates, so as to achieve the purpose of enhancing calculation accuracy and efficiency. The simulation results show the feasibility and effectiveness of the proposed algorithm.