The invention relates to a failure diagnosis method of power equipment, comprising the following steps: (1) installing a failure monitoring and predicting system comprising a reciprocating engine, an oil extraction device, an oil preprocessor, an oil data analyzing and storing module, a failure diagnosis module, a magnetoelectric sensor, a voltage sensor, a pressure sensor, a temperature sensor, a vibration sensor, an ultrasonic wave sensor, a signal conditioning box, an analog-to-digital conversion module and a data acquisition module; (2) judging whether the acquired data exceed the upper limit of the threshold values of the normal phase and the amplitude according to the oil analysis data and the data acquired by each sensor, if not, continuing operating the equipment, if so, carrying out phase deviation and amplitude data quantifying; (3) establishing a fuzzy evaluation matrix of the amplitude and the phase signal according to a fuzzy evaluation method; and (4) carrying out failure diagnosis on various acquired data according to the fuzzy evaluation matrix. The invention has overall acquisition means, can satisfy the failure predicting and monitoring functions and has high diagnosis precision. The invention can be widely applied to various reciprocating machines.