The yield of methanol is heavily influenced by the composition of the syngas which varies from waste type and gasification conditions. This study uniquely explores machine learning (ML) and artificial neural network (ANN) algorithms with shapley additive values (SHAP) analysis to predict syngas compositions and interpret the complexities of waste gasification. Aspen Plus is then utilized to investigate the impact of syngas compositions on the economics of methanol production. Among the different ML models, gradient boosting regression achieved remarkable accuracy in predicting carbon monoxide and carbon dioxide, while ANN excelled in predicting nitrogen and hydrogen based on the coefficient of determination (>0.95). The feature importance analysis demonstrated the gasification agent type as the most important feature. Steam gasification of waste sawdust, straw, and wood chips showed the highest methanol yield. Economic analysis indicated the lowest levelized cost of methanol (LCMeOH) to be 1195.79 USD/MT which further decreased to 933.97 USD/MT upon hydrogen addition. Uncertainty analysis using Monte Carlo method showed the possible minimum LCMeOH to be 254.05 USD/MT. This study provides a comprehensive picture of the economic viability of waste-to-methanol production and provides an important reference for stakeholders on a global scale to quickly assess the feasibility of waste-to-energy projects.