<p>Conventional machine learning algorithms are primarily designed for vector-based data and hence often fail to exploit the inherent continuity and smoothness that characterize functional data. These approaches treat functional data as a discrete and constrained sequence of observations, thereby overlooking crucial insights into the continuous and coherent functional patterns that define the data’s underlying generative process. It also suffers from issues associated with highly correlated measurements within each functional object. To overcome these limitations, it is imperative to develop models that can effectively utilize both the functional form and derivative information of the data. In this paper, we propose a functional twin support vector machine (FTWSVM) that aims at generating two nonparallel hyperplanes such that each hyperplane is closer to one of the two classes and is as far as possible from the other for functional data. We have used derivative information along with functional information to obtain the classifiers and a voting strategy is applied to compute classification results. By integrating derivative information and employing a voting-based classification strategy, the proposed FTWSVM enhances discriminative capability and robustness. Experiments conducted on benchmark datasets prove the efficacy of our FTWSVM in binary as well as multi-class settings.</p>