TreeMortNet is a Python package for predicting individual tree-level mortality using NAIP imagery and lidar-derived tree crown polygons. It uses a pre-trained residual convolutional neural network to make predictions, and should be applicable to most forest and woodland types of the Western US.