The invention discloses a photovoltaic inverter DC-side resistance parameter fluctuation coefficient prediction method. the method comprises steps: (1) according to parameters acquired through real-time monitoring, a photovoltaic inverter DC-side resistance parameter fluctuation coefficient evolution mathematical model time sequence is built; (2) according to an optimized genetic-Bayesian algorithm, the evolution mathematical model time sequence is processed; and (3) according to a prediction value, the photovoltaic inverter DC-side resistance parameter fluctuation coefficient is calculated. The technical problems of evaluation accuracy and not high photovoltaic utilization rate are solved. The method has the technical effects that the photovoltaic inverter evaluation accuracy is improved; the problem of electric energy quality of a distribution network due to photovoltaic power station access can be avoided; the photovoltaic utilization rate is improved; the reliability of the power grid power system is improved; and the economy of the distribution network power system in distributed photovoltaics is improved.