This research is intended to produce a smart sensor system using a Neural Network which is implemented through the design of an Occupation monitoring tool and application (presence). Sensor equipment uses Arduino UNO, EST8266 Wifi, DHT22 and LDR. The readings of these sensors are then sent remotely via the internet and read by the application. Based on the research training conducted, a prediction model will be born where this model will translate the sensor data received and then generate data on the presence in an environment where the sensor is placed. IoT sensor tools and applications that are placed in several locations are very useful in carrying out daily human activities such as monitoring security in the home environment, schools, offices and warehouses for storing goods.