Gasification, among the other thermochemical processes, stands as a promising method for sustainable waste management while generating valuable H2-rich syngas. In the last few years, there has been a significant surge in the adoption of machine learning (ML) techniques for modelling biomass as well as gasification and pyrolysis of waste. The present study unveiled the supremacy of Gaussian Process Regression (GPR) algorithm as compared to other ML models, in predicting syngas yield [Ysyn (Nm3/kg of biomass)] and carbon conversion efficiency[ηCarbon(%)] respectively, with a high prediction accuracy of R2 > 0.9 in both the cases. The necessary datasets were collected from experimental trials, at reaction conditions of temperature (T): 550–1000 °C, air flow rate (AFR): 0.02–0.04 L/min, and reaction time of 2 h, in a two-stage fixed bed reactor, which were identified on the Basis of thermodynamic as well as stoichiometric analysis. Additionally, the models were employed in Bayesian optimization within a weighted multi-objective framework to obtain the optimal operational parameters, focusing on maximizing both yield and conversion simultaneously. Finally, this result was validated experimentally to check the sanctity of the data-driven optimization.