With the widely application of the Multi-platform and Multi-sensor in the Air Combat Rivalry Evaluation, the accomplishment of the better data Pre-processing of Multi-platform Air Combat Rivalry Evaluation will play a vital role in route selection、 track fusion、 evaluation and prediction of real -time and afterwards air combat training. Through building multi-platform air combat training evaluation system, the measuring data from varied platforms is analyzed. The Least Square method and coordinate transformation are adopted to realize the time and space registration between varied platforms, the process accumulated error is analyzed and disposed based on the Gaussian Sum Estimate Kalman Filter algorithm. Simulation analysis is done, the result shows that the matched curve between measurement track and real track is good in fitting result after filter processing, and the purpose of perfecting the Pre-processing of multi-platform air combat rivalry evaluation is achieved.