The integration of machine learning (ML) into boiler control systems presents a promising venture for enhancing operational efficiency and decreasing emission production of domestic heat sources. This study explores the application of MATLAB-based machine learning models and strategies in smart boiler control, focusing on their capacity to optimize performance and reduce environmental impact. The research centers on two key improvements: implementing a variable-speed air ventilation control and a variable adaptive speed fuel feeder for biomass boilers. By analyzing various ML algorithms and their implementation in MATLAB, the study provides an overview of current methodologies. The findings demonstrate curve fitting techniques and model training that predict thermal output. Additionally, the study highlights the advantages and challenges associated with deploying MATLAB-based ML models. Future work will involve refining these control systems via a IoT experimental setup with continuous fan and fuel feed system using machine learning methods and exploring their application to further enhance sustainability in the domestic heating sector.