Abstract: In order to provide new technical support for the variable temperature drying process and realize the adaptive control of drying temperature based on the change of moisture content, the study designed a temperature adaptive control system with the function of online detection of material moisture content. A moisture content online detection model with weight detection value, air velocity, the temperature of load sensor elastic substrate, airflow impingement distance as inputs, and the real material weight as outputs was established by using a convolutional neural network. A validation test of the moisture content online detection model was carried out. The results showed that the model meets the accuracy requirements of online moisture content detection in the variable temperature drying process, and the coefficient of determination R² and root mean square error (RMSE) of the five groups of model validation tests were 0.9934 and 1.20% in that order. In this reasearch, an improved neural network-PID (INN-PID) controller was designed to realize temperature control in the variable temperature drying process. The dynamic performance of PID, neural network-PID (NN-PID), and INN-PID controllers was simulated in MATLAB software with unit step signal as input. The three controllers were tested for drying temperature control at 50-55 ℃. The results showed that the control stability and regulation time of the INN-PID controller were significantly better than the other two controllers in the simulation test, the drying temperature control results had approximately the same law with the simulation results, and the peak time of the INN-PID controller was 208.00 s, the regulation time was 120.59 s, and the maximum overshooting was 4.87%, which meets the requirements of temperature control in the variable-temperature drying process. In this paper, a temperature adaptive control system was built in the air-impingement dryer, and the temperature adaptive control test based on the moisture content change was carried out. The results showed that the system could quickly and effectively regulate the drying temperature in the variable temperature drying process based on the change in moisture content. This research is of great significance for improving the automation level of drying equipment, developing new variable-temperature drying processes, and providing reference for multi-information fusion detection and control strategy research in other fields.