Nowadays, with the continuous improvement of the level of urbanization, there are more and more empty nesters, and the home safety of empty nesters has become a social focus. With the frequent reporting of the murder of women living alone, the home safety of women living alone has also aroused widespread concern in the society. As a key part of the home security system, the access control system plays an important role in ensuring the home safety of empty nesters and women living alone. At present, the access control security measures on the market are mainly supported by two independent devices, remote monitoring and intelligent door locks, and do not play a coupling role in authenticating visitors and pushing remote alarms, and do not do statistical modeling and analysis of visitor data. This makes the function of remote monitoring and intelligent door lock completely independent, can not play the intelligent role of danger warning and prediction in advance, and can not really improve the safety factor of the home access control system to ensure home safety. In addition, due to the limited architecture, it is relatively difficult to replace or add firmware such as camera sensors for smart door locks. If you want to experience new functions or improve the recognition accuracy, you need to replace the equipment, which leads to poor user experience, unnecessary waste of resources, and does not conform to the theme of sustainable development. The system uses strawberry pie as a hardware development platform, Android as a software development platform, using today's popular face recognition and cloud interaction technology, while the system has built-in access to view, access statistics, white list editing, a key call for help and other functions. The hardware side uses Python as a development language to achieve face detection and control of the sensor. The software side uses the MVVP framework to achieve efficient development of APP. The server side uses Tencent cloud platform to realize face recognition, and uses SpringBoot framework as the back end to realize remote interaction by deploying to Aliyun server. The intelligent Internet of Things access control system based on the identity authentication and data statistical modeling of visitors in the video environment has been developed, which realizes the remote early warning of unfamiliar visitors and the analysis of visitor data, and realizes the double guarantee of access control security in terms of timely alarm and advance prediction.