Fintech, or financial technology, has advanced the financial industry significantly by using cutting-edge technologies like blockchain and artificial intelligence (AI) to improve and automate services. The study emphasizes how well support vector machines (SVM) can secure financial systems. This study investigates the use of the Legendre kernel in a support vector machine (SVM) framework for fraud detection in vehicle insurance data. By using orthogonal polynomials as basis functions, we attempt to capture nonlinear relationships in the dataset that a traditional linear kernel might miss. This research helps illustrate the potential of orthogonal polynomial kernels for fraud detection in machine-learning models.