Abstract: In order to segment the target information of the microscopic image of Chinese herbal medicine (CHM) effectively, an improved Pulse Coupled Neural Networks (PCNN) automatic segmentation method of CHM microscopic image was proposed based on Gray Wolf Optimization (GWO) algorithm. Firstly, the traditional PCNN model was simplified and improved from the perspective of adapting to the processing of microscopic images; Secondly, Shannon entropy was extracted from the training image as the fitness function of GWO to adaptively adjust the key parameter (link coefficient β) of PCNN, then the optimal segmentation of image target was realized; Finally, the proposed algorithm was compared with clustering segmentation, OTSU and traditional PCNN method, and the four processing methods were objectively evaluated by common medical image segmentation evaluation criteria such as Dice coefficient, volume overlap error, relative volume, accuracy and intersection union ratio. The experimental results showed that, the proposed method could realize the adaptive segmentation of image, better maintain the information of image detail, texture and edge, had high segmentation accuracy for different microscopic images, could improve the segmentation performance of image and had strong applicability.