The complexity of the spatial environment and the ambiguity of the radar signal processing lower the reliability of the single radar in the precision tracking system. It has become a trend of using the radar in conjunction with other measurement equipment to improve the measurement accuracy, and this involves the comprehensive and multi-dimensional data fusion processing of the multi-sensor data. This paper summarized and analyzed the existing mechanisms and application methods of multi-sensor data fusion, and categorized the traditional and the new fusion methods according to their application scope, advantages, and disadvantages. These mainly include radar target detection, tracking, and recognition, as well as various heterogeneous sensors in the radar data fusion. Moreover, this paper also compared the data fusion algorithms and methods of different multi-sensors.