To make full use of the advantages of the data measured both by a radar and by an infrared sensor, a fusion method for radar and infrared data is proposed. On the basis of least square theory, the radar data and the redundant information of an infrared sensor when it measures at a high frequency before and after the radar measures are used. The optimal estimation of a target is obtained by incorporating two weighting factors: time dimension and measurement dimension. The simulation test result shows that compared with the equality weighting factor fusion method, the fusion data information obtained by the proposed method can reflect the motion state of a target better and can provide more accurate observation value for the subsequent target tracking.