The invention discloses a particle analysis system. The system includes a sample test module used for uniformly covering a test piece of a microfluidic chip with a mixed liquid; an image acquisition module used for obtaining a digital image of a blood sample in the test piece; an image preprocessing module used for performing particle region enhancement on a grayscale image, and performing particle region coarse segmentation by utilizing a maximum interclass variance threshold method; a feature point constructing module used for judging an aggregated particle region and a single particle region according to an optimal global threshold, and separately performing area peak feature extraction on the aggregated particle region and the single particle region; and a recognition and counting module used for taking a set of particle feature region peak feature points as input, recognizing the particle region by utilizing a MeanShift clustering algorithm, and performing particle counting according to a recognition result. Through adoption of the particle analysis system, aggregated particles with insignificant features of regions and chain codes; impurity interference in the sample can be accurately removed; and particle counting analysis is more accurate and reliable.