The production mode of hand-sewn for complex shape multi-layer thermal insulation (MLI) of satellite has seriously affected its production efficiency and quality. Automatic sewing technology is an important means to improve MLI production efficiency. First, the problem of automatic creation of sewing patterns and paths needs to be solved, but there is no targeted research on MLI with special sewing requirements. In this paper, the principles of the sewing path are determined by combing the boundary conditions. Based on the chain code histogram and its spatial distribution entropy, the internal opening of MLI can be quickly identified to determine the nodes through which the sewing path passes. Then, the simulated annealing algorithm which can quickly converge to the global optimal solution is used to solve the NP-hard problem and realize the generation of the sewing path. For the special sewing requirements, the Laguerre diagram-based A-Star algorithm is used to realize the avoidance sewing path to the internal opening, and the MLI contour reinforcement sewing path planning is abstracted into the inward quantitative offset line generation problem of the contour. So the problem of automatic generation for sewing path of complex shape MLI can be solved.