The paper aims to design a temperature control system by integrating drosophila optimization algorithm and radial basis neural network to improve the temperature control accuracy of salt packaging line. The structure of the control system was introduced. The self-learning and adaptive ability of RBF neural network were used to realize the online adjustment of PID controller parameters, which can ensure the adaptive control of sealing temperature. The initial value of neural network was optimized by drosophila optimization algorithm, and the global searching ability of neural network was improved. Finally, the simulation and experimental analysis were carried out. The results showed that the temperature deviation can be controlled below 1%, and the control algorithm had good stability. It took less time to reach the stable state, and the overshoot of the system was significantly reduced, which improved the accuracy and stability of sealing temperature control to a certain extent. The control system has ideal control performance and can meet the requirements of controlling the sealing temperature of edible salt packaging.