In order to improve the assembly quality of the automobile tail light area, on the basis of Gaussian image and mathematical statistics, taking the Hausdorff distance for reference, this paper quantified the matching status on Gaussian scale space, and established the objective function and the subsidiary verification target parameter for optimization, to find the best matching adjustment method by computer program. Based on the verification and analysis of the validity of the theory above, the test on the application of actual engineering data shows that it is effectual for improving the assembly quality with strong operational.