The invention provides a puccinia striiformis f. sp. tritici uredospore automatic statistical counting method, which can be used for quantitative monitoring and accurate counting of puccinia striiformis f. sp. tritici uredospores in outdoor field air. The method comprises the following steps of collecting a microscopic image of the puccinia striiformis f. sp. tritici uredospores to a computer through a TPBZ3 spore capturer and a research-level microscope; and then performing a K-means clustering algorithm, morphological processing, shape factor and area-based adhered spore judgment, concavity-based adhered contour segmentation and adhered spore contour section fusion and counting, wherein the counting accuracy can reach 98.6%, and automatic statistics and counting of the uredospores are realized. The method is high in counting precision and high in practicality, solves the difficult problem, urgently needed to be solved, of counting of the puccinia striiformis f. sp. tritici uredospores in the outdoor field air, and provides technical support for online development of a puccinia striiformis f. sp. tritici uredospore monitoring device.