Modeling of photovoltaic(PV)system output is the premise of evaluation of its capacity value and planning. For such modeling

how to deal with stochastic and fluctuant nature of solar irradiation is the key point. Based on the theory of cluster analysis

the output model of PV system is established from the data mining perspective and the model is applied in the reliability evaluation of power system incorporating PV generation. Firstly

the conventional fuzzy C-means(FCM)algorithm is improved to solve the problem of initial iterative value sensitivity. Secondly

the improved FCM algorithm is used to make a cluster analysis on historical solar irradiation data. The number of clusters and the clustering centers are determined. Then

based on the membership matrix and Monte-Carlo simulation(MCS)method

the simulation method of solar irradiation is proposed. Finally

through the IEEE-RTS system

the method proposed is used to evaluate the generation system reliability. The simulation results have demonstrated that the proposed model and method are effective and highly efficient.