The worldwide network has to a great extent perceived that the Earth's atmosphere is evolved a lot in the last few decades. In particular, the Climatic Action attempts to both reduction of European Union (EU) ozone harming substance outflows and enhance the efficiency by decreasing the consumption of essential energy. In Retail shops and commercial buildings are liable to consistently monitor and control for the Heating, Ventilation, and Air Conditioning (HVAC) and refrigeration systems. As per the monstrous Internet Of thigh's (IoT) collection of data, there are unnecessary utilization of energy may happen because of manual activity in the Retail shops and commercial buildings. Recent decades, smart supermarkets are implemented by tuning HVAC systems and the refrigeration system automatically for the purpose of improving the satisfaction of customers and also optimizing energy consumption. To achieve an agenda, in this paper, it plans to build up a technique for (1) investigating the sensor model depending on the detected data; (2) constructing a forecasting model for a working status of systems, and proposing a Firefly based optimized Long Short-Term Memory Network (FOLSTM) model for the advance forecasting of data; (3) improving the forecast precision utilizing FOLSTM with the comparison of conventional methods. This FOLSTM technique with real-time collected data from the sensors of an HVAC and the refrigeration framework where the information is appropriate to general IoT hardware for investigating the accuracy and the determining prediction status.