<p>An evaporator is a surface that transfers heat and vaporizes a volatile liquid refrigerant by drawing heat away from a product or refrigerated area. It is commonly referred to as a chiller, freezer, or cooler and is positioned on the low-pressure side of the refrigeration system, between the expansion device and the compressor. Its primary function is to absorb heat from the material or area that needs to be cooled using refrigerant. Energy assessments can help identify the feasibility and potential benefits of adopting alternative refrigerants within the existing refrigeration system. In this paper, an efficient method for extracting features and classifying using Improved Physics Informed Neural Network (IPINN) for determine the cooling capacity of the evaporator in the Refrigerator system is developed. Recorded data such as pressure, electricity consumption, time, energy consumption, and temperature are used as inputs, which undergo pre-processing to convert raw data into meaningful information. Pre-processing methods like Min–Max normalization and the Gaussian process for MI are utilized to standardize and replace the missing values in a dataset. After pre-processing data are fed into the feature extraction technique called Forward Search Informed Team Work Optimization (FSITWO) to select the important data optimally. Finally, the Improved Physics Informed Neural Network (IPINN) algorithm-based classifier is used to determine the evaporator’s cooling capability. The proposed method achieves an accuracy of 95.23%, a Positive Predictive Value (PPV) of 89.31%, a True Positive Rate (TPR) of 89%, and an error rate of 4.77%. Thus this proposed work is the better choice for identifying the cooling capacity of an evaporator.</p>