Particle filtering algorithm has been widely used in solving nonlinear/non-Gaussian filtering problems. In this paper, a new particle filter is proposed, which is based on the unscented Kalman filter (UKF) and the extended Kalman filter (EKF), and takes a divide-and-conquer sampling strategy. It first uses a mixed Kalman filter, which combines UKF and EKF, as proposal distribution to generate part of the particles, and then uses the transition prior for another part. The experiment results show that this new particle filter can reduce time cost in addition to giving higher accuracy compared to other particle filters.