The separation process in the oil and gas industry occurs in a three-phase separator. The separator needs to be controlled so that the oil production is stable and gets maximum results. In this study, an adaptive PID control system was designed using a neural network, which will adjust the PID adaptively during the separation process. This control system can be controlled remotely because it is designed with a combination of the Wireless Sensor and Actuator Network (WSAN) systems using Long Range (LoRa) modules and Internet of Things (IoT) modules that are integrated with a cloud database system. The system was designed using IEEE 802.15.4g and the HTTP protocol. The results showed that the WSAN communication system with LoRa devices can communicate in real-time at 300 meters with an average delay time of 145ms without pocket loss. When combined with adaptive level control systems using Neural Network-PID (NN-PID), the adaptive level control performs better than the PID Controller. At mixed room level and oil room level, NN-PID is more responsive compared to PID control. NN-PID can dampen oscillations better than PID. However, both controls have the same steady state error. The proposed WSAN design can work well over a long distance and is also stable and reliable in performance.