In view of bearings-only target tracking problem, traditional linearization algorithm can not meet the requirements of nonlinear, non-Gaussian and real-time target tracking. An improved sampling-importancere-sampling filter(SIRF) algorithm was proposed for highly non-linear bearing-only tracking system where the common particle filters often fail to catch and keep tracking of the emitter. Specific improvements are to remove normalization step and use non-normalized weights directly. While maintaining high accuracy estimation ability at the same time, the algorithm has strong robustness, is an effective solution to state estimation of nonlinear system. Finally, experiments verify that the tracking effect of the improved SIRF algorithm is better than the Gaussian particle filter (GPF) algorithm.