A tractor is a typical representative of power machinery in agricultural production. Two working conditions of tractors include road transportation and field operation. These working conditions can also be subdivided into plowing, rotary tillage, and fertilization, according to the carried agricultural implements. However, there are high oil consumption and low energy-saving in tractors, where the external load fluctuates frequently, as the working environment changes. Furthermore, a large number of gears have been equipped and frequently shifted to fully meet the needs of different working conditions in the traditional tractors, leading to the difficulty of tractor operation. In this study, a hybrid-mechanical-hydraulic power system was designed for a tractor using continuously variable transmission and hybrid power technology. The driving and transmission modes of the system were realized to obtain the speed regulation curve of the continuously variable transmission. The mounted hydro-mechanical continuously variable transmission (HMCVT) was also optimized for the best performance of the power system. Taking the engine economy curve as the target, the economic speed ratio map of HMCVT was drawn by the MATLAB platform. The target speed ratio was then obtained, according to the vehicle speed and throttle opening. As such, a speed ratio control strategy was formulated using the engine economy curve. Moreover, an energy management strategy was constructed using the minimum equivalent fuel consumption. The equivalent factor was integrated into the working division using the logic threshold. The penalty function was then introduced to maintain the battery State of Charge (SOC) stability. A strategy was designed for the rule-based mode division and the minimum fuel consumption power allocation using the adaptive equivalent factor. Specifically, the vehicle speed, demand torque, battery SOC, and speed ratio were firstly used as the logic threshold for the mode division, and then the minimum instantaneous equivalent fuel consumption was used as the objective function. Finally, the corresponding constraint conditions were introduced to calculate the minimum fuel for the engine and motor at any moment. The dynamic model of the tractor system was established in the SimulationX software. A test bench was also built using a dynamometer. The simulation and experimental analysis of the tractor were performed on the three typical working conditions of plowing, harvesting, and transportation. The results showed that the error between the simulation and test value was less than 5%, indicating a reliable model. The HMCVT efficiency was above 0.80 under the three working conditions, whereas, the efficiency of the whole power system was around 0.4. The SOC values were achieved at +1.96, -0.37, and -0.19 at the end of the plowing, harvesting, and transportation, respectively, indicating near the target value. The fuel consumption values were 2.72, 6.80, and 1.77 L in the plowing, harvesting, and transportation, respectively. Therefore, the 18% and 15% oil consumption was saved under the power output condition, compared with the power shift tractor, and the continuously variable transmission tract published by the German Agricultural Association, respectively. By contrast, 15% and 19% of the oil were saved under the transportation condition, compared with the power shift tractor, and the continuously variable transmission tractor, respectively. The 9%-20% fuel was also saved under three working conditions, indicating the feasible power system and the control strategy. The finding can provide a better solution to reduce the tractor energy consumption under multi-working conditions.