This study uses data-driven decision-making and deep learning approaches to forecast and model the possible dangers of soil and groundwater contamination based on an existing database of industrial facilities. The project attempts to establish and validate information and norms concerning soil contamination concerns. The research focuses on the gas station industry due to concerns about data integrity, availability, and the clarity of existing targets inside the database. It examines data on gas station monitoring, administration, and setup to determine the possibility for soil and groundwater pollution. The study creates a deep learning model and assesses its performance measures, concluding that the binary classification model passes performance requirements, with a confusion matrix suggesting 92.11% accuracy, 91.21% precision, and 93.21% recall. The area under the ROC curve is 0.897, indicating strong performance. An examination of the relative importance of the model parameters reveals that the top six factors influencing the potential risk of soil and groundwater contamination are, in order, contamination potential, municipality, pipeline protection, historical average contamination potential, tank age, and tank material.