<p>Composite materials have revolutionized material science by integrating physical and chemical properties to enhance performance. Machine and Deep learning techniques surfaced as advanced resources in predicting the tensile strength of composite materials by overcoming the challenges of traditional approaches. This study presents a novel multi-class classification framework to predict the tensile strength of composite materials (low, medium, and high classes) using a diverse set of real-world manufacturing and structural features including matrix-filler ratios, resin requirements, stitching geometry, and curing compositions. Various machine learning classifiers such as decision tree, gradient boosting, support vector machine, logistic regression, random forest etc., along with some neural network models are trained and examined based on multiple metrics. While evaluating machine learning classifiers, support vector machine obtains the highest accuracy of 70% with the best values of precision, recall as 1.00, and F1 score followed by Naïve Bayes and logistic regression with 69% accuracy each. On the other hand, Gradient Boosting Model generates 86% accuracy where for the high strength class it has a 0.82, 0.91, and 0.87 as precision, recall, and F1 score respectively. For the low-strength class, it achieved precision (0.88), recall (0.93), and F1 score (0.90). The medium-strength class showed slightly lower performance, with recall, precision, along with F1 score as 0.73, 0.87, and 0.79 respectively. The findings of the paper emphasize the prospective of using different learning classifiers as a viable technique for predicting the tensile strength of composite materials by providing a cost-effective and efficient solution.</p>