Abstract: An effective supply mode, solar power is gradually gaining much attention for the environmental friendliness and convenience. Among them, a photovoltaic/thermal co-generation technology is usually utilized to improve the power generation efficiency of photovoltaic cells, as the operating temperature rises. As such, the redundant heat generated by the photovoltaic cells was reused as the heat source for the temperature difference of the power generation system to realize secondary power generation. Moreover, the Maximum Power Point Tracking (MPPT) control is also required to achieve the optimal potential of the co-generation system. In this study, a new MPPT control of photovoltaic/temperature difference was proposed further to combine the constant voltage and hyperbolic tangent type adaptive variable step size, in response to the oscillation and misjudgment caused by the fixed step size of traditional conductance increment. Two advantages were included here: First, the control was the fast tracking to the area near the nonlinear region of the maximum power point using 0.78 times of the system open-circuit voltage, suitable for the great changing environmental conditions. Second, the step size was adjusted adaptively and quickly, according to the change of external environmental conditions, when the MPPT was tracking to the nonlinear region near the maximum power points. For instance, the light intensity was used to reduce the system oscillation, indicating the monotonic increase and fast variation in the hyperbolic tangent function. Furthermore, a simulation model was established to evaluate the performance of adaptive variable step conductance increment in the MPPT control of a combined photovoltaic/thermal power generation system using the Matlab/Simulink software. Specifically, Jinao JAMG-6-60-250/SI photovoltaic module was set as the photovoltaic cell model, and Xinghe F40550 was the thermoelectric chip model. Simulation results show that the step changes were consistent under the drastic variation in the light intensity, while the response speed was obviously improved with the rapid adjustment for tracking the maximum power point. At the same time, the step size was kept at 0, after the output power of He system was stabilized. There were only small fluctuations and errors in the steady-state output power, indicating that the MPPT control performed well. Correspondingly, an MPPT hardware experiment was conducted to further verify the feasibility at Northeast Agricultural University in Harbin in October 2020. Two periods A (8:00-9:00) and B (12:00-13:00) were selected, when the illumination and temperature were gradually enhanced to remain unchanged. The hardware experimental results show that the system was quickly tracked and stabilized at the maximum power point within 15ms, where the steady-state error was less than 0.3%, indicating more robust to external environmental disturbances and higher energy utilization. Consequently, an excellent balance was achieved in the system response speed and steady-state accuracy. The finding can provide a promising potential to the implementation of hardware, such as digital signal processors in practice.