Due to the methane emission from the mining of coal and industrial byproducts including CFCs, depletion of the ozone layer causes raised global warming and worsens climate change. One chlorine atom can lead to the destruction of 1,00,000 ozone molecules causing depletion at a much faster rate than natural replacement. The research studies the impact of industrial expansion of Delhi on ozone depletion through the development of hybrid predictive modeling by the integration of Convolutional Neural Networks (CNN) and Bi-directional Long Short-Term Memory (Bi-LSTM) networks. While the CNN extracts important spatial features, the Bi-LSTM captures temporal dependencies, thus achieving precise forecasts. To further improve the extraction of relevant data from encoded sequences, a multi-head attention layer is placed between encoder layers. The model performance was assessed using Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Normalized Root Mean Square Error (NRMSE). The simulation results indicate that the CNN-Bi-LSTM model is characterized by a MAE of 0.214, an RMSE of 0.268, and a MAPE of 34.22%, and esteems it to be better than the traditional models, such as LSTM, SVR, and Random Forest. The model predicts ozone depletion for short and long periods of time, thus ensuring accurate future projections and reliable monitoring. The developed system was tested in various conditions for 2100 hours and found to be accurate, reliable, and robust. These findings would indicate the suitability of the system for real-time monitoring and forecast at an appropriate time for policy intervention and recommendation to minimize further depletion.