Information fusion technology is one of the key technologies of millimeter wave/infrared (MMW/IR) combined guidance. Its target tracking algorithm can directly affect the performance of a MMW/IR combined guidance system. According to the target tracking of MMW/IR combined guidance, the measurement data from both a MMW sensor and an IR sensor are fused firstly and an improved alpha-beta tracking filter algorithm is proposed. The algorithm can obtain the filter gain in light of the maneuver situation of a target in real time and adjust the filter equations in time, so as to obtain good tracking performance. Finally, the linear motion and the course-changed linear motion of the target are simulated with the improved algorithm. The simulation result shows that compared with other filter algorithms, the improved algorithm has a better tracking effectiveness, a higher tracking accuracy and a less computation quantity.