This paper introduces a new, metaheuristic optimization algorithm, named an Improved Metaheuristic Equilibrium Optimizer (IMEO). The algorithm Equilibrium Optimizer (EO), is inspired by its method of estimating both equilibrium and dynamics, based on mass balance models. Studying the EO closely, we find that EO does not have the potential to get closer to the optimal global solution when it solves certain problems. To improve EO efficiency, this paper suggests using an improvement, called an elite opposition learning-based, that increase in the speed and accuracy of EO convergence, and helps the algorithm to get a better solutions. Falling into local optima is a big problem, EO suffers from the fact that when we look deeply at the standard EO mathematical formula, we found that the algorithm is trying to get out of the local optima, but sometimes it can't, so we're introducing a new mathematical formula based on using cosine trigonometric function. To validate the proposed algorithm efficiency, The IMEO algorithm is evaluated on 23 benchmarks and compared with other common naturalistic heuristic algorithms. The statistical analysis shows that the results of IMEO achieve better performance compared to the standard EO and several well-known algorithms on several benchmark issues