The reciprocating progressive scanning is an effective way to increase the imaging frame rate of laser confocal scanning microscopic imaging. However, serious image dislocations during image reconstruction are introduced as the frame rate increases. According to the motion characteristics of a high-speed galvanometer, the dislocation principle of reconstructed images by the laser confocal high-speed scanning microscopic imaging system is analyzed. A dislocation evaluation algorithm of reconstructed images is designed based on the morphological gradients. In addition, the single-objective constrained particle swarm algorithm is used to realize the dislocation correction of reconstructed images by combining with the search for the minimum point of dislocation evaluation. The experimental results show that the proposed algorithm is suitable for the dislocation correction of reconstructed images even when the imaging frame is up to 300 frame/s. Compared with that of the original images without dislocation correction, the imaging resolution after correction increases by 68.83%. Moreover, this algorithm is also suitable for the image reconstruction under different mirror combinations and different scanning frame rates.