Developing intelligent analytical tools requires pre-processing data and finding relevant features that best reinforce the performance of the predictive algorithms. Feature selection plays a significant role in maximizing the accuracy of machine learning algorithms since the presence of redundant and irrelevant attributes deteriorates the performance of the learning process and increases its complexity. Feature selection is a combinatorial optimization problem that can be formulated as a multi-objective optimization problem with the purpose of maximizing the classification performance and minimizing the number of irrelevant features. It is considered an NP hard optimization problem since having a number of (n) features produces a large search space of size () of different permutations of features. An eminent type of optimizer for tackling such an exhausting search process is evolutionary, which mimic evolutionary processes in nature to solve problems in computers. Salp Swarm Algorithm (SSA) is a well-established metaheuristic that was inspired by the foraging behavior of salps in deep oceans and has proved to be beneficial in estimating global optima for optimization problems. The objective of this article is to promote and boost the performance of the multi-objective SSA for feature selection. Therefore, it proposes an enhanced multi-objective SSA algorithm (MODSSA-lbest) that adopts two essential components: the dynamic time-varying strategy and local fittest solutions. These components assist the SSA algorithm in balancing exploration and exploitation. Thus, it converges faster while avoiding locally optimal solutions. The proposed approach (MODSSA-lbest) is tested on 13 benchmark datasets and compared with the well-regarded Multi-Objective Evolutionary Algorithms (MOEAs). The results show that the MODSSA-lbest achieves significantly promising results versus its counterpart algorithms.