<p>Bio-composite materials are increasingly recognized for their significant role in enhancing building energy efficiency. Their use helps in reducing the carbon footprint of buildings and lead to lower energy consumption for heating and cooling due to their lower thermal conductivities. This property is of particular importance in construction, where effective thermal insulation is critical to achieving energy efficiency. The traditional methods to study thermal conductivity have relied on theoretical models, numerical simulation, or various experimental methods. In this study, we employ another approach to predict the thermal conductivity of bio-composite materials using machine learning techniques. With a set of experimental data available in the literature, we explored the effectiveness of seven machine learning algorithms: Linear Regression (LR), Ridge Regression (RR), Lasso Regression (LSR), Polynomial Regression (PR), Support Vector Regression (SVR), and Random Forest (RF) in predicting the thermal conductivity of bio-composite materials. We utilized Spearman and Pearson’s coefficient correlation analysis to identify key factors influencing thermal conductivity. Through the evaluation of three performance metrics, and assessing the impact of constituent densities on prediction quality, we found that Random Forest model excelled, achieving optimal prediction metrics (<i>R</i>² = 0.926, MSE = 0.001, and RMSE = 0.034); these results even outperformed theoretical models namely Maxwell, beck and Woodside models. This research demonstrates that machine learning techniques provide a powerful tool for predicting thermal conductivity of bio-composite materials.</p>