In order to sovle the problem of target tracking under the complex scenes,this paper proposes an adaptive unscented particle filter(A-UPF) algorithm for the system with unknown noise.The new algorithm estimates and corrects the statistic characteristics of the system unknown noise in real-time by improved Sage-Husa estimator,and produces optimal distribution function with unscented Kalman filter.The new algorithm reduces the estimation error effectively and improves the anti-noise ability of the system.The experimental results show that the method proposed in this paper has high precision and strong robustness for target tracking under the complicated conditions.