This study presents a machine learning-based investigation of the co-gasification of biomass and coal, focusing on predictive modeling of syngas yield. Experimental data sets from published works were utilized to capture a wide range of operational conditions and feedstock compositions, overcoming the constraints of resource-intensive experimental campaigns. Three models, Linear Regression (LR), Tweedie, and Extreme Gradient Boosting (XGBoost) were developed, and tested to see how well they worked. LR gave a clear baseline with modest predictive accuracy, but XGBoost did far better than LR. It was better at capturing nonlinear linkages and complicated interactions that are common in co-gasification processes. With minimal prediction errors and an unbiased residual distribution, the XGBoost model got training and testing R2 values of 1 and 0.9199, respectively. These results show that sophisticated machine learning approaches are robust and can be used to optimize and regulate co-gasification operations on a large scale. The study sheds light on the most important process factors and suggests using ML-driven methods to improve modelling accuracy beyond what is possible with traditional methods.