Abstract: To address the large inertia, time delay, and nonlinear characteristics of refrigeration systems in refrigerated vehicles for less-than-truckload (LTL) cold chain logistics, as well as temperature fluctuations caused by frequent compartment door openings during multi-stop urban distribution, this study developed a composite temperature control system to achieve high-precision adaptive temperature control and ensure the quality of fresh food transportation.Based on the principles of energy conservation and heat transfer, a dynamic heat exchange model of the refrigerated compartment was established, which comprehensively considered heat leakage through the enclosure, respiration heat generated by cargo, and thermal disturbances caused by compartment door openings. An Improved Sparrow Search-Genetic Algorithm (ISSA-GA) was proposed. The algorithm combines the strong global exploration capability of the Sparrow Search Algorithm (SSA) with the advantages of crossover and mutation operators in the Genetic Algorithm (GA) for discrete variable optimization. The introduction of the t-distribution mutation strategy enhanced the global search capability, while the dynamic-radius perturbation mechanism enabled a smooth transition between exploration and exploitation phases, thereby achieving efficient collaborative optimization of fuzzy PID membership functions and rule bases.Furthermore, a feedforward-feedback composite control strategy was designed to monitor compartment door status in real time and trigger feedforward compensation. The convergence curves demonstrated that ISSA-GA exhibited superior global search capability. The standard GA stagnated after the 45th generation, whereas the proposed hybrid algorithm achieved global convergence at the 73rd generation. Based on the Integral of Time-weighted Absolute Error (ITAE) criterion, the final fitness value was improved by 18.02 % compared with GA and by 1.95 % compared with the standard SSA.The proposed method was validated through MATLAB/Simulink simulations and experiments using a 50 L scaled-down experimental chamber. The results showed that the settling time of the ISSA-GAFuzzy controller was 60.5 s, which was reduced by 62.54 % and 59.40 % compared with the conventional PID and fuzzy PID controllers, respectively. The maximum overshoot was 0.69 %, representing reductions of 94.90 % and 93.02 % compared with the conventional PID and fuzzy PID controllers, respectively. After introducing feedforward compensation, the temperature fluctuation caused by compartment door opening was reduced from 2.26 ℃ to 0.98 ℃. Continuous temperature setpoint variation experiments demonstrated that the proposed system achieved smooth transitions under both small and large setpoint variations, with negligible overshoot and rapid stabilization characteristics. Experimental verification under setpoints of 5 ℃ and 0 ℃ showed that the temperature tracking rates reached 83.2 % and 85.14 %, respectively. The maximum temperature error was maintained within 0.26 ℃, exhibiting a low steady-state fluctuation coefficient. Door-opening disturbance rejection experiments further verified the effectiveness of the feedforward compensation mechanism. The feedforward-enhanced ISSA-GAFuzzy controller achieved a disturbance recovery time of 110.80 s, which was 43.61 % shorter than that without feedforward compensation. Meanwhile, the temperature rise amplitude was reduced by 42.86 %. The results indicate that the proposed temperature control system exhibits significant advantages in response speed, steady-state control accuracy, and disturbance rejection robustness. It can effectively adapt to the complex operating conditions of LTL cold chain logistics and provide a reliable technical solution for fresh food cold chain transportation.