Abstract: In order to optimize the startup program of compressor sets in a compressor station of gas pipeline, a mixed-integer nonlinear programming model was developed, and two calculation methods were proposed. In the first method, the objective function is dealt by means of piecewise linear interpolation and constraints are linearized, so that the model is transformed into a mixed-integer linear programming model. Then, it is solved by the MILP solver. And in the second method, the model is transformed into a multistage decision problem, which is solved by means of dynamic programming method. Case calculation shows that both calculation methods are satisfactory. The dynamic programming method runs faster when there are fewer compressor sets in the compressor station, but as the number of compressor sets increases, the calculation amount increases approximately exponentially. The research results provide the reference to work out startup program of compressor stations with different types of compressor sets.