Accurate prediction of energy consumption, particularly in the natural gas sector, plays a crucial role in strategic policymaking, resource allocation, and achieving sustainable development. This study aims to enhance the accuracy of annual natural gas consumption forecasting in Iran by developing a hybrid model based on the Grey Markov model combined with optimization techniques. In the proposed approach, the parameters of the Grey Markov model are optimally tuned using the Simulated Annealing algorithm to improve forecasting precision. The data utilized were extracted from Iran’s energy balance sheets covering the period from 2002 to 2021, which were divided into training and testing sets. The performance of the proposed model was evaluated against traditional Grey and Grey Markov models for forecasting the country’s natural gas consumption up to the year 2035. Results indicate that the proposed model significantly outperforms the baseline models, achieving a reduction in mean squared error by 28.64 units and a decrease in mean absolute percentage error by 1.52%. The model provides a more accurate and stable forecast, delivering a logical and conservative long-term consumption trend that avoids unrealistic growth projections. Key parameter reductions, such as the number of Markov states and optimized coefficient tuning, contributed to improved convergence and reduced fluctuations. Therefore, this model can serve as a reliable and practical framework for forecasting natural gas demand, aiding in addressing energy imbalances.