The two-phase impulse turbine (TPIT) can improve the efficiency and economy of various thermal systems, because it can directly convert the internal energy of thermal fluids into the shaft power of the turbine with no needs of additional low-temperature sub-systems. However, current efficiency of reported TPITs is much lower than gas or water turbines, due to unknown characteristics of the two-phase flow in TPITs. This paper comprehensively analyzed the performance characteristics of a TPIT, which was used in the refrigeration system. CFD methods were employed to analyze the two-phase flow based on flashing models and validated with experimental results. Three average methods were used to derive the Euler power and compared with the output power derived by the torque on each blade, while the average method based on the modified factor can evaluate the Euler power accurately. The non-uniform flow caused by the nozzle position was significant, while the liquid film near pressure side and the shroud were also three dimensional and affected by wall effects. The liquid layer on the shroud was related to the negative torque on blade tips. Results presented in this paper are constructive for designing efficient thermal systems without additional energy recovery devices.