In order to understand drone modeling and PID control, the author proposes a research on drone modeling and PID control based on machine vision. The author first analyzed an effective method for stabilizing the flight attitude of unmanned aerial vehicles. Based on the forces and motion principles of vertical, left, right, forward and backward movements, and yaw movements of multi rotor unmanned aerial vehicles, explore the construction of a mathematical model, select a suitable reference coordinate system, and choose the quaternion method as the mathematical method to describe the attitude of the four rotor unmanned aerial vehicle. When describing the motion mode of the multi rotor unmanned aerial vehicle, use roll angle, pitch angle, and yaw angle to verify the relationship, and serve as input for subsequent attitude control. Reasoning the change matrix of the transformation between the body coordinate system and the navigation coordinate system, establishing a nonlinear dynamic model. Based on the complementary filtering algorithm and the extended Kalman filtering algorithm, the collected data is fused and estimated. Combined with a cascade PID controller, it serves as a framework, and intelligent fuzzy control theory is added internally. The three values Kp, Ki, and Kd of the PID parameters can be adjusted in real-time, making the drone more robust, when encountering unexpected situations, flight control can intelligently adjust parameters to quickly stabilize the drone. The simulation results using MATLAB/Simulink show that the optimized fuzzy control PID parameters can effectively adjust the attitude angle rate and control the position response rate of the drone, thereby improving the stability of the drone.