The invention discloses a photovoltaic module fault diagnosis method based on external characteristic electrical parameters, comprising the following steps: first, carrying out feature extraction on the external characteristic electrical parameters of a photovoltaic module under different faults; then, using a BP neural network, a RBF neural network, an Elman neural network and a RVM algorithm to establish four photovoltaic module fault diagnosis models based on external characteristic electrical parameters to carry out preliminary fault diagnosis of the photovoltaic module; and finally, putting forward a photovoltaic module data fusion fault diagnosis model based on improved evidence similarity, taking the diagnosis result of the four models as the basic probability assignment (BPA) function value of the improved data fusion algorithm, and carrying out fusion diagnosis output at a decision-making layer. Through the photovoltaic module fault diagnosis method based on external characteristic electrical parameters, the defect that the use of a single fault diagnosis method results in low reliability is solved.