In order to solve the problem that the real-time performance of the laser cutting visual identity system was poor in the recognition of large and multi-target images, and the recognition rate of targets with deflection angle was low or even zero, niche genetic algorithm was used to study the large-surface multi-target matching recognition algorithm. Theoretical analysis and experimental verification were carried out. The results show that, whether the target is rotating or not, the algorithm can reach recognition rate of 100%. The computation time of the algorithm is 5 times~10 times faster than that of the traditional algorithm. The algorithm has good practicability in identifying multiple targets with rotation angle.