The rapid development of robotics technology has led to its increasingly widespread application in various complex scenarios. However, these scenes are usually dynamic and subject to effects such as uneven lighting and occlusion, which can affect the positioning accuracy of robots. Therefore, a research on position self calibration of intelligent robots in complex scenes based on laser SLAM is proposed. Using coverage grid map and posterior probability to construct a robot movement map, in order to ensure the accuracy of robot position self calibration results, particle filtering algorithm is introduced. The position state sample set is obtained through importance sampling principle, and the prior probability of robot position state is obtained based on the position state transition function and sensor observation principle. Bayesian formula is used to update the prior probability and obtain the posterior probability. Monte Carlo simulation is used to The Dirac function updates the position state weights, uses resampling to remove the degraded parts of the position state weights, and obtains the optimal calibration result through multiple iterations. The experimental results show that laser SLAM can achieve intelligent robot position self calibration in complex scenes, with small positioning error and fast convergence speed.