Feature selection is a very important step in machine learning and data mining, as it aims at identifying and preserving only relevant features that contribute a lot of significance to the task of prediction or classification, irrelevant and redundant aspects are eliminated at the same time. A novel approach, Opposition-Based Whale Optimization (WOA-OB) algorithm applied to feature selection is proposed to improve the performance for the classification task as well as its reliability. The Whale Optimization Algorithm, which is the latest metaheuristic optimization method under the larger nature-inspired algorithms category generates new populations at random during the exploration and exploitation phases, as other population-based algorithms do. As a result, it may be inclined to deviate from the ideal solution or be stuck in a local optimum. WOA-OB is an extension of the Whale Optimization Algorithm which makes use of the opposition-based method to improve the efficiency of the algorithm. WOA-OB looks into solutions that are opposite to original values to see whether going in the other direction yields a better outcome. After conducting a thorough comparison, it was concluded that the WOA-OB is more efficient in the sense that it is able to give more accurate results. The approach was evaluated on 4 different datasets. The cervical cancer dataset gave the best accuracy of 100% while the glass dataset had the highest accuracy of 98.46% which shows the efficacy of the proposed method.