Multi CNN has recently gained popularity in image classification applications. In particular, Computer vision has acquired a lot of attraction due to its numerous potential uses in food quality management. Among all the dry fruits available in India, the cashew nut is a significant crop. Specifically high-quality cashew nuts are quite popular on the worldwide market. Although there are a variety of approaches for automatically identifying cashew nuts, the majority of them concentrate on a single view image of the cashew nut. The fundamental issue with current methods for recognizing whole and split cashew nuts is that a single view image of a cashew nut cannot encompass the entire view of a cashew nut, resulting in low classification accuracy. We proposed Multi-view CNN to provide a novel framework for classifying three types of cashew nuts. Images of the sample cashew nuts are taken from three distinct angles (top, left, and right) and fed into the proposed modified CNN architecture. For categorization, the modified CNN extracts and combines many elements from these three images and obtains the accuracy of 98.87%.