The working efficiency of cotton-picking machine is always influenced by its working speed. To solve this problem, an algorithm for real-time adjusting the working speed was put forward based on fuzzy PID control technology. The walking speed regulation model was also established to achieve the optimal control of the working speed of cotton-picking machine. The algorithm combined the traditional PID control with fuzzy algorithm by acquiring the real-time parameters under the input load gradient congestion conditions, such as cotton production online. Both the fuzzy rules and reasoning method of PID were used to set online parameters and adjust the real-time working speed to achieve optimal control of operating speed. The field test was conducted by using 4MZ—5 five-line cotton-picking machine based on FPGA hardware. The testing result showed that the algorithm response speed was stable in 75s. According to the cotton volume of online production, several parameters were adjusted by using fuzzy rules under the load gradient and block boundary conditions, which ensured the stability of picking up for cotton-picking machine with the best speed and improved working efficiency. It also reduced the loss in the process of picking cotton, and effectively avoided the tube plugging failure as a result of large amount of cotton. The system showed a strong flexibility and versatility and it could be used to reduce the driver’s workload and improve the automation and information level for large equipment.