A fuel cell has come up as an upcoming new and renewable energy recently. The fuel cell is the system which is converting chemical energy to power and heat. When applied to fuel cell in buildings, PEMFC(Polymer electrolyte membrane fuel cell) is mainly used due to relatively low operating temperature (50-100℃). Performance prediction models for fuel cell are actively being studied, however, researches are mainly focused on analysis about stack or electrochemical reaction in the fuel cell. The purpose of this study is to suggest empirical performance prediction model for PEMFC by statistically analyzing PEMFC operation data and to evaluate the feasibility of empirically derived models. Also, combining the data from this and previous study, electrical efficiency and thermal efficiency model are derived to completement each other. Fuel cell performance is predicted under any operating condition using these models.