In order to obtain the optimal minimum variance state estimation for electro-optical tracking,a nonlinear Kalman filter algorithm is provided by combining reduced state Kalman filter and first-order linear in nonlinear system,and the algorithm structure is summarized in detail. Then,it is applied in nonlinear measurement electro-optical tracking system and compared the performances of reduced state Kalman filter with extended Kalman filter and unscented Kalman filter. The Matlab simulation results show that combining reduced state Kalman filter and first-order linear in nonlinear system is valid,and the performance outperforms the extended Kalman filter and unscented Kalman filter.