Toxicity assessment is a crucial aspect of drug development, evaluating the harm a compound may inflict on an organism, notably within organ systems like the liver. This study employs the Camel Algorithm for feature selection and the Support Vector Machine (SVM) method, specifically targeting NR-AhR toxicity. Utilizing the Tox21 Data Challenge dataset, a comprehensive exploration of three SVM kernel func- tions— Linear, Radial Basis Function (RBF), and Polynomial—is conducted, accompanied by thorough hyperparameter tuning. The results showcase improvements across all kernels, with the RBF kernel emerging as the most effective. The optimized model, integrating the Camel Algorithm and the RBF kernel in SVM, surpasses alternative approaches, demonstrating exceptional pre- dictive capabilities. Upon evaluation with test data, this refined model achieves an impressive accuracy of 0.921 and an F1-Score of 0.612. In summary, this research not only contributes to the ongoing enhancement of methodologies for toxicity prediction but also presents a robust approach within the NR-AhR dataset context. The findings underscore the significance of the Camel Algorithm and SVM in advancing safer and more effective pharmaceutical development, marking a significant stride in the field.