In cloud computing, task scheduling plays a major role and the efficient schedule of tasks can increase the cloud system efficiency. To successfully meet the dynamic requirements of end-users’ applications, advanced scheduling techniques should be in place to ensure optimal mapping of tasks to cloud resources. In this paper, a modified Henry gas solubility optimization (HGSO) is presented which is based on the whale optimization algorithm (WOA) and a comprehensive opposition-based learning (COBL) for optimum task scheduling. The proposed method is named HGSWC. In the proposed HGSWC, WOA is utilized as a local search procedure in order to improve the quality of solutions, whereas COBL is employed to improve the worst solutions by computing their opposite solutions and then selecting the best among them. HGSWC is validated on a set of thirty-six optimization benchmark functions, and it is contrasted with conventional HGSO and WOA. The proposed HGSWC has been proved to perform better than the comparison algorithms. Moreover, the performance of HGSWC has also been tested on a set of synthetic and real workloads including fifteen different task scheduling problems. The results obtained through simulation experiments demonstrate that HGSWC finds near optimal solutions with no computational overhead as well as outperforms six well-known metaheuristic algorithms.