Air-fuel mixture combustion produces toxic gases that result from combustion. In this research, the goal is to reduce the environmental impact of gas turbine emissions during the power generation process. By using the Predictive Emission Monitoring System, machinery emissions can be predicted, leading to a reduction in these emissions. This study uses a hybrid model, combining Convolutional Neural Networks (CNN) and Bidirectional Long-Short-Term Memory (Bi-LSTM) to forecast these emissions. It is the CNN component's tasks to extract features from the multi-dimensional emission data. By contrast, the Bi-LSTM model, which is renowned for its ability to handle sequential data and long-term dependencies, manages prediction and classification using these sequences. By combining CNN-BiLSTM attention regression with emissions data, the hybrid model is able to capture both spatial and sequential characteristics, improving prediction accuracy.A promising result can be seen when evaluating this model's effectiveness for CO and NOx emissions. A slight underestimation bias of -0.01 is associated with CO predictions with an RMSE of 0.064, MAE of 0.04, and R2 of 0.82. NOx predictions have an RMSE of 0.051, MAE of 0.036, R2 of 0.887, and a slight overestimation bias of +0.01. In essence, this study presents a promising way to predict and reduce gas turbine emissions by using a hybrid model. There is room for refinement, especially when it comes to CO emission predictions based on the results of the model.