A diagnosis method of rotating machinery fault based on genetic programming and weighted evidence theory is proposed to improve accuracy of fault diagnosis.Multi-fault characteristic parameters are extracted by using the genetic programming and the preliminarily diagnose fault of rotating machinery is realized.Then their diagnosis is separately used to act as evidences.The diagnostic accuracy of initial diagnosis is regarded as the weight of the evidences.Finally

all evidences are aggregated and analyzed according to the combined rules to reduce the uncertainty of fault diagnosis

and fault diagnosis is fulfilled accurately.Experimental result shows that the proposed method can improve the accuracy of fault diagnosis.