Classification of facial skin types is important in facial care. This research aims to develop a facial skin type classification system using the Convolutional Neural Network (CNN) method with the ResNet-50 architecture. This research uses a dataset of 1,119 facial skin images with 3 classes: normal, dry, and oily. The research stages include: data pre-processing in the form of image resizing and normalization, model training using ResNet-50, and data testing. The research results show an accuracy of 0.9986, a loss of 0.0040, and a high F1-Score. However, unbalanced data distribution causes overfitting, so future research needs to use a more balanced dataset.