The study aimed to classify cotton water stress using Convolutional Neural Networks. The cotton image was collected by a UAV at Lubbock, Texas. A DJI Phantom 4 was utilized as the UAV platform for collecting high-resolution RGB images at an altitude of 90 meters. I have built a CNN model to classify the dataset into four irrigation treatments making use of Resnet. This was my final project for ENGR 689: Machine Learning for Computer Vision.