The photovoltaic-thermoelectric coupled system has drawn widespread attention in recent years due to its potential of utilization of full-spectrum solar energy. The output power of thermoelectric (TE) is usually much less than that of photovoltaic (PV), because the temperature difference of TE module is usually very small. Thus, the optical concentrating method has been used to increase the temperature difference, but it increases complexity of the structure and extra cost for the sunlight tracking system. Another practical way is using the thermal concentrating method to increase the heat flux of the TE module. However, for there two PV-TE coupled systems i.e. optical and thermal concentrating systems, there exists a confrontation between the PV cell and the TE module. The efficiency of PV cell decreases with the increasing temperature, while the performance of the TE module increases with the increasing temperature difference. In some conditions, the total efficiency of the PV-TE coupled system may be even lower than that without TE module. Therefore, it is necessary to investigate the optimal combination mode of PV cell and TE module for the coupled system. In this paper, a thermodynamic model based on energy-balanced equation is established to investigate the structural optimization and efficiency improvement of PV-TE coupled power generation system, which considers the effects of the thermal concentration factor, the length of thermocouples, the loading resistance of TE module and thermal diffusion factor. The influence rules of these factors on power generation efficiency of PV-TE coupled system without and with concentrator are also theoretically analyzed. (1) For the PV-TE coupled system without concentrator, the influences of thermal concentration factor, the length of thermocouples and the ratio of loading resistance to the internal resistance of TE module on the total efficiency of PV-TE coupled system are analyzed under four efficiency temperature coefficients of PV cells ( β ref =0.001, 0.002, 0.003, 0.004 K−1). It is found that, when the figure of merit of TE module Z is 4×10−3 K−1 and β ref is 0.001 K−1, the total efficiency increases initially and then decreases with the increases of the thermal concentration factor, the length of thermocouples and the ratio of loading resistance to the internal resistance of TE module. Therefore, there exist optimal values of thermal concentration factor, length of thermocouples and loading resistance of TE module to maximize the total efficiency. (2) For the PV-TE coupled system with concentrator, the influence of thermal diffusion factor on the total efficiency of PV-TE coupled system under five efficiency temperature coefficients of PV cells ( β ref=0.001, 0.0015, 0.002, 0.0025, 0.003 K−1) is studied for the figure of merit of TE module Z =2.545×10−3 K−1 and Z = 4×10−3 K−1. The results shown that, for the cases of Z =2.545×10−3 K−1, β ref= 0.001 K−1; Z = 4×10−3 K−1, β ref= 0.001 K−1; Z = 4×10−3 K−1, β ref= 0.0015 K−1, the total efficiency decreases as the thermal diffusion factor increases. While for other cases, the total efficiency increases significantly as the thermal diffusion factor increases. It has also shown that the increases of efficiency caused by the increase in thermal diffusion factor becomes small when the thermal diffusion factor is over 5.12. The reason is that the heat transfer becomes fast when the thermal diffusion factor becomes larger, which leads to the increase of temperature of the PV cell and the hot side of the TE becomes small as the thermal diffusion factor is over 5.12. Based on the above results, it is suggested that an optimal thermal diffusion around 3.0 can have a better balance of both cost and total efficiency of the PV-TE coupled system with concentrator.