<p>This study develops a sustainable and economical production system for deteriorating green products with variable demand, which depends on the green level, warranty period, and selling price of the items. In this system, the production process is imperfect, and the rate of imperfection is known and constant. A green-level-dependent unit production cost is considered. Moreover, to provide free service to customers, the warranty cost per unit is modelled as an increasing function of the product’s warranty period. Two models (Model-1 and Model-2) are developed within the proposed system. While preservation technology is not utilised in Model-1, it is applied in Model-2 to slow down the rate of deterioration. The optimisation problems involve highly nonlinear objective functions due to the incorporation of realistic nonlinear assumptions. Therefore, these problems cannot be solved using analytical methods. The Equilibrium Optimiser Algorithm (EOA) is employed to solve the optimisation problems. Additionally, the same problems are solved using the Artificial Electric Field Algorithm (AEFA), Zebra Optimisation Algorithm (ZOA), Whale Optimisation Algorithm (WOA), and Grey Wolf Optimiser Algorithm (GWOA), and the obtained results are compared with those from EOA.Two numerical examples are considered to assess the feasibility and practical relevance of the proposed system. The results indicate that the average profit of Model-2 is higher than that of Model-1, demonstrating that Model-2 is more profitable. Furthermore, it is observed that the Equilibrium Optimiser Algorithm (EOA) outperforms the Artificial Electric Field Algorithm (AEFA), Grey Wolf Optimiser Algorithm (GWOA), Whale Optimisation Algorithm (WOA), and Zebra Optimisation Algorithm (ZOA) in solving the given problems. Finally, sensitivity analyses are conducted to examine the impact of key parameters on the optimal policy, and from the results, several managerial insights are derived.</p>