Abstract: The appearance of multi-processor and multi-mode scheduling problems marks a development trend that becomes more complicated and flexibility-oriented. Through recursive microstructure method, the thesis offers a uniform definition for the current scheduling problem as well as some complex problems that are hard to be clarified with traditional methods. In terms of principal-subordinate structure, general flexibility degree (GFD) is used here as an indicator to measure the scheduling flexibility, and GFD classification provides a ground for linear and integer programming solution to the problem. Quantitative experiment not only verifies that appropriate flexibility can effectively optimize the best solution of the scheduling problem, but also results in a statistical conclusion towards the reasonable machine group size.