Concrete is an essential building material that finds extensive use in a wide range of applications, particularly in the field of structural engineering where its flexural strength is of utmost importance. Over the past few years, the incorporation of machine learning (ML) models has demonstrated great potential in accurately predicting and enhancing concrete performance. The primary objective of this study is to forecast the flexural strength of concrete that incorporates recycled aggregates. The dataset utilized in this study comprises 302 data points, encompassing crucial input parameters such as Cement, Sand, Natural Fine Aggregate, Recycled Fine Aggregate, Natural Coarse Aggregate, Recycled Coarse Aggregate, Water, W/C ratio, and Superplasticizer, whereas, output parameter is the flexural strength of concrete (fck). Two ML models, the Adaptive Boosting Regressor (Adaboost) and the Extreme Gradient Boosting Regressor (XGB), are utilized to predict strength. In addition, a range of data visualization techniques which include scatter plots and histograms along with errors, coefficient of determination (R2), and others are employed. According to the findings of the study, XGB demonstrates superior performance compared to Adaboost, as evidenced by an R2 value of 0.83 for XGB, while Adaboost achieved a value of 0.59. It is evident that XGB demonstrates a superior ability to capture the fluctuations in flexural strength and closely align with the observed data points. The results highlight the promise of ML models, specifically XGB, in effectively forecasting the flexural strength of concrete that includes recycled aggregates. This research contributes to the progress of sustainable construction methods.