Industrial groundwater consumption in Balasore, Odisha, has a significant impact on the region’s groundwater level. Studies in the area have shown that the rapid industrialization in Odisha, particularly in Balasore, is contributing to a decline in the groundwater level due to continuous water usage without adequate conservation measures. Therefore, an effective groundwater level forecasting method is essential for resource management and environmental sustainability. However, traditional methods are unable to capture the complexity that includes the influence of industrial consumption and meteorology. This study examines the impact of the HIL Limited industry, taking into consideration several predictors like temperature, rainfall, daylight duration, shortwave radiation, and industrial water consumption, in Balasore, Odisha. During this research, different machine learning and deep learning models such as Multiple Linear Regression (MLR), Random Forest Regressor (RFR), XGBoost Regressor (XGBR), Artificial Neural Networks (ANN), Gated Recurrent Unit Networks (GRU), and Long Short-Term Memory networks (LSTM) were trained and tested. The ANNs exhibited higher model performance with R2 = 0.9800, RMSE = 0.1582, MAE = 0.1007, and MAPE = 1.1140. Moreover, the ANNs offer better performance predictions due to their improved capability of representing the complex non-linear dynamics of groundwater levels. This paper also includes visual representations to compare the models for performance metrics and understand the groundwater dynamics.