In chemical process optimization, identifying conditions that balance production rate and thermal risks is crucial. This paper presents a surrogate-assisted optimization methodology that integrates parameters uncertainty, specifically focusing on synthesizing γ-valerolactone (GVL) in adiabatic and batch modes. A surrogate model was established to elucidate the relationships between input variables, production rate and risk index, which reduces the computational burden associated with complex differential equations. The Latin Hypercube Sampling method was employed to assess how uncertainties propagate through the processes. This study formulates a multi-objective optimization model that seeks to find a balance between the highest possible GVL production rate and the lowest probability of failure under deterministic and uncertain scenarios. The results in Pareto charts illustrate the possible operating conditions and determine the optimized initial conditions. This approach serves as a model for optimizing complex chemical processes, balancing production capacity and safety while considering uncertainty management.