As a market-based mechanism which can actively promote energy efficiency, Energy Performance Contracting (EPC) is now universally accepted throughout China and has been incorporated into various business models. Four of these business models are frequently adopted in China: the Share Savings Model, the Guaranteed Savings Model, the Energy-cost Trust Model and the Finance Lease Model. The choice of a suitable business model has been recognized as one of the most important factors in the effective implementation of EPC projects. In this paper, the selection of EPC business models is formulated as a multi-criteria decision making (MCDM) problem. When selecting an EPC business model, it is natural for decision-makers (DMs) to express their preferences/opinions in linguistic terms. Given the characteristics of EPC projects, taking DMs' psychological behavior into account is of considerable and realistic significance. TODIM (an acronym in Portuguese of interactive and multicriteria decision making) has been verified as an effective behavioral decision-making method. To address the problem of EPC business model selection, an extended TODIM method based on multi-granularity linguistic terms and entropy measure is proposed. The sensitivity and comparison analyses show that DMs' psychological behavior can affect the selection order with respect to the given EPC business models. One case of a real-life campus EPC project demonstrates that the validity and applicability of the proposed extended TODIM method for solving the EPC business model selection problem.