<p>In the twenty-first century, emissions of carbon dioxide have become a serious concern. The world's socioeconomic issues are significantly impacted by growing global average temperature and its effects on climate change. Controlling this carbon emission is urgently needed. This paper introduces a novel model for predicting the CO₂ emission based on copula entropy And jackal optimisation techniques for feature selection methods. A dataset with 76 features for CO₂ emission is used in this paper. After applying the one-hot encoding to convert categorical variables into a numerical representation that machine learning models can utilise, the features after one-hot encoding become 117 features. As the number of features increased, the computation time is also increased. Feature selection will solve this problem. The fundamental idea in this paper is to evaluate the relevance And reliance of features using copula entropy by reducing the dimensionality of high-dimensional feature selection problems, and then to use the golden jackal optimisation method to determine which subset of characteristics is the best. The features minimised to 63 features with the proposed feature selection method. The initial CO₂ prediction from the dataset, which includes 117 features, was generated using a linear regression model. The results indicated a mean squared error (MSE) of 4.85 and an R-squared (R²) value of 0.68. These results are not acceptable as we need to improve the prediction accuracy. To address this, another prediction was conducted using a Linear regression model, utilising the 63 features selected from the proposed GJOA-CE. The results indicate that R-squared (R²) increased to 0.9998. The results confirm that the proposed model can effectively predict the CO₂ emissions with small number of features.</p>