The invention relates to a pest image recognition method based on a parallel convolution neural network. Compared with the prior art, the method solves the problems of low pest recognition rate and poor robustness, wherein the problems are caused by less pest image samples. The method comprises the steps that a training image is collected and preprocessed; a parallel convolution neural network model is constructed and trained; an image to be measured is collected and preprocessed; and a test sample is input into the trained parallel convolution neural network model to realize automatic recognition of the pest image. According to the invention, through the constructed parallel convolution neural network, network training is adequate in the case of insufficient number of samples, and the feature of stronger classification ability plays a greater role.