This paper proposes a fusion of multiplexed optical sensors and variable integral PID methods for intelligent handling robots to deal with their inability to realize accurate trajectories in complex environments. First， a short-wavelength semiconductor laser sensor and multiple grayscale sensors were used to sample the trajectory in real time， and the collected data were preprocessed by applying the sliding average filtering algorithm. Then， the deviation of the data from the equilibrium state was calculated using a combined multi-optical sensing and variable integral PID control algorithm. The accumulation and fusion were weighted， which serves as the input of the variable integral PID control algorithm and reduces the convergence distance to 10 cm. Finally， to expand the applicability of the gripping structure of the object block， a four degrees-of-freedom robotic arm， extractor pump， and load table were used to form a grasping structure. The experimental results show that the algorithm can perform the functions of QR code recognition， stable trajectory， and block grasping and handling in 3 min， in both indoor and outdoor environments. The robot has a task handling accuracy rate of 100%. The algorithm improved the stability and environmental adaptability of the trajectory of the handling robot， strengthened the real-time feedback of the robot route in the process of handling the trajectory， and accelerated the robot's response to the trajectory error， which is of great significance for automation technology.