A transformer fault diagnosis method based on RIMER（belief rule-base inference methodology using the evidential reasoning approach）expert system and data of dissolved gasses analysis（DGA）is proposed to deal with the complex nonlinear relation in transformer fault diagnosis. Considering the probability uncertainty and fuzzy uncertainty of the transformer fault characters

a new DGA fault diagnosis model is established based on the evidence consequence and the parameters of belief rule base from both IEC three-ratio method and training of fault samples of dissolved gasses. The RIMER-DGA transformer fault diagnosis model solves the problem of fault code missing in IEC three-ratio method

thus improves the fault diagnosis accuracy and obtains better description of mixed-type fault. Simulation shows that the proposed method is simple and feasible.