Abstract:Support Vector Machine(SVM) is a machine learning technique demonstrating many peculiar advantages in solving machine learning problems of small sample,nonlinear and high dimensional.It is applicable to the field of function prediction,pattern recognition and data classification.The algorithm is applied to operation optimization,clean production,and fault diagnosis in thermal power plant,and its parameter prediction accuracy can satisfy the engineering applications,and thns provides a new research direction for the thermal power plant operation optimization and fault diagnosis.