In order to better understand and research the face detection based on convolutional neural network(CNN), this paper analyzed the main difficulties faced by face detection and its solution in low image quality face detection and the performance optimization of face detection system. This paper summarizes the development process and direction of face detection technology based on convolutional neural network since 2015. It covers occlusion, scale, small face cluster, speed, precision, and multi-task region proposal network face detection in face detection tasks. A summary statement is made to predict the future development of face detection based on convolutional neural networks.