The necessity for accurate and dependable concrete compressive strength projections is critical to the structural integrity and longevity of building projects. Unlike traditional methods, this work looks into using machine learning algorithms to predict concrete’s compressive strength in an effort to effort to increase efficiency and accuracy. The goal of the project is to apply mean squared error (MSE) techniques to optimize the prediction models. We go over the value of accurate predictions of compressive strength in the construction industry as well as the limitations of utilizing conventional methods. The study investigates how successfully these algorithms detect complex patterns in the data and adapt to different mix ratios, curing times compositions of concrete. We discover that the mean squared error is a trustworthy statistic to gauge prediction accuracy, which aids in solving the issue statistic of assessing model efficiency. The study emphasizes how important mean square error reduction is to achieving optimal prediction performance. Machine learning model optimization techniques that are to like feature engineering and hyperparameter tuning are discussed in order to decrease the mean square error (MSE).