Abstract: In order to quantitatively predict the change of moisture content in the drying process of tangerine peel heat pump, the image features in the drying process of tangerine peel were extracted based on image processing technology, and the moisture content prediction model was established. Tangerine peel images in different drying periods were collected, image processing was used to preprocess the tangerine peel images, and a total of 12 image features including 6 color features and 6 texture features were extracted. The relationship between feature parameters and moisture content changes was analyzed, the relevant image features were taken as the input of the model, and the moisture content of tangerine peel was taken as the output of the model. The prediction models of tangerine peel drying moisture content based on BP neural network and support vector machine were respectively established for comparative analysis, and the best prediction models of moisture content in different drying periods were obtained. The results show that the prediction effect of support vector machine is better, with an accuracy of 99.01% and a mean square error of 0.006 5. The model runs stably, and the prediction result of moisture content is accurate and fast, which can provide a scientific basis for the online prediction of moisture content in the drying process of tangerine peel.