In the oil industry, multiphase separators are of paramount importance in oil production, as they have a significant impact on overall oil production capacity. However, the occurrence of slug flow, which is characterized by undesirable behavior, can cause extensive damage to the surface separator and pipelines. Consequently, the development of intelligent separators and the implementation of effective flow regime control mechanisms are crucial. To address this issue, our research aimed to investigate the behavior of different flow regimes entering the separator and evaluate the intelligence system of the separator. Initially, we constructed a laboratory-scale multiphase flow loop as a platform to examine various flow regimes entering the separator. Through laboratory tests employing a PID controller, we collected data concerning the liquid level, gas pressure, input flow rate, and control signals of the separator. Subsequently, we accurately modeled the laboratory system and conducted simulations in a multiple-input-multiple-output mode with model predictive control. After ensuring the effective operation of the laboratory smart control system, we integrated its control model with a production system containing a real fractured reservoir. The results demonstrated that the implementation of the smart integrated production system could enhance oil production by approximately 53% over a 20-year simulation period.