Feature selection is a key process in machine learning that reduces data dimensionality without losing significant information. This paper presents multi-objective with genetic programming-based feature selection for obtaining optimal feature subsets with good classification accuracy. The method is tested on eight benchmark datasets, where the objective is to reduce the number of features and improve predictive performance. Unlike the conventional approaches in which a single objective is highlighted, this approach balances feature reduction and accuracy improvement in an optimal way. As observed from the experimental results, the proposed approach reduces the features significantly without losing accuracy in reaching or exceeding existing models. Through improved interpretability and computational expense, this approach is beneficial in handling high-dimensional datasets. The paper showcases the potential of evolutionary algorithms for feature selection, which benefits machine learning model designing to create better and scalable machine learning models.