<p>A feasible method for predicting carbon emissions may help in the development of a carbon minimization strategy, which is crucial in a CO₂ constrained environment to avoid global warming. India is one of the growing economies that has led to higher carbon emissions as an effect of its reliance on fossil fuels. Keeping in this issue, this study is to suggest an improved machine learning method called Adaptive Neuro-Fuzzy Inference System-Snake Optimizer (ANFIS-SO) learning to predict and analyse the interactions between energy use, agriculture, land use, industrial processes, waste, and CO₂ emissions in India. The proposed model is evaluated by using the database of Indian emissions data (World Bank database) from the year 2000 to the year 2021. The performance of the ANFIS-SO technique is compared to the ANFIS, ANFIS-CESBAS, and ANFIS-RSA models using known statistical benchmarks. From experimental results, the ANFIS-SO model outperforms other models with respect to the statistical benchmarks, such as mean absolute percentage error (0.0136), root mean squared error (0.0329), root mean square relative error (0.0183), mean absolute error (0.0209), and the determination coefficient (0.9919). Moreover, the trained ANFIS-SO model is used to estimate India's prospective CO₂ emissions up to the year 2040. This proposed prediction model could be a powerful tool for the system for managing the atmosphere, which is very unstable, complicated, and non-linear. India can apply these prediction strategies to minimize carbon emissions to prevent a unified policy that restricts economic growth in impoverished areas.</p> Graphical Abstract <p></p>