Aiming at the problem that traditional forecast models can’t get satisfactory results, support vector machine (SVM) optimized by genetic algorithm (GA) was built as a forecast model for maintenance cost of aviation equipment. The model combined GA and SVM, which used GA to optimize the parameters of SVM. Through the example analysis and comparison, results show that the model has more accurate results and extensibility than SVR, BP network, partial least squares regression and multivariate linear regression in the forecast of maintenance cost of aviation equipment.