This paper presents a proposal for the analysis of the variable of water stress of an avocado plantation, in the town of Mollepata-Cusco, Peru. The methodology considers the acquisition of multispectral images of medium-scale avocado plantations using a multispectral image installed in UAV. The images obtained were pre-processed using the median filter and histogram equalization, both methods highlight the characteristics of the images obtained in the NIR band and RED band. From the NIR and RED bands, the image of the NDVI index is found, then the segmentation is performed using the Otsu multi-Thresholding technique, the values obtained are used for classification. All this process was carried out using MATLAB software and the results were compared with commercial PIX4D software. The results show that in the avocado tree (strong avocado), water stress manifests itself concentrically from the edges towards the stem, it is also observed that the avocado tree, has high resistance to water stress. The study of these images shows the NDVI indices as a function of water stress in avocado plantations, contributing to farmers to make decisions to improve irrigation and differentiated treatment.