The support vector machine technology is applied to study the fault diagnosis for supercharge system of diesel engine. Based on the different features under various faults of supercharge system of diesel engine and the signal variables of the fault of the system’s components, dimensionless daters can be gotten after normalization processing. The utilization of SVM with good generation ability wad adopted to establish the diagnosis model for supercharge system of diesel engine, which can make the system failure diagnosis through inputting the diagnosis data. The average accuracy is 99.8% in twenty tests. The diagnosing results are consistent with the factual results.