Optoelectronic and radar data fusion can achieve complementary of two single sensors measuring information, and improve target recognition and tracking capability. Aiming at the presence of outliers in localization of combined sensor system to dynamic motion target, meanwhile, in order to solve a single sensor tracking filtering divergence problem, a resist outliers interacting multiple model unscented Kalman filter (IMM-UKF) fusion algorithm is proposed. Under the condition of two coordinate radar providing target range and azimuth angle, parameter solving model is established and the pitch angle of target is obtained. Fusion filter is carried on by combining angle information provided by photoelectric sensor. Experiment and simulation results show that the algorithm can effectively integrate radar and optoelectronic measurement data, eliminate the interference of outliers, suppress filter divergence and improve positioning accuracy.