In this study, we introduce an advanced design for an advertisement deployment device, distinguished by its utilization of multi-sensor data fusion techniques coupled with an LSTM-based recommendation system algorithm. This apparatus is adept at receiving and processing real-time user behavior and environmental data from an array of sensors. Through meticulous data preprocessing, feature extraction, and LSTM recommendation strategies, we can dynamically deliver advertisements that closely align with a user's behavior and environment. In essence, this design not only optimizes the precision of advertisement deployment but also enhances personalized user interaction experiences, offering significant academic and practical value.