The tasks scheduling problem is about mapping tasks to resources so that the objectives related to both users and service providers are satisfied. In this paper, three improvements are presented to enhance the ability of the Red Fox Optimization (RFO) algorithm. Firstly, a Quasi-Opposition Based Learning method is employed for generating the initial population and a Levy flight method is used to enhance the exploration ability of newly generated foxes. Secondly, two fuzzy control systems are used to provide a balance between exploration and exploitation activities. Thirdly, a cornu-spiral movement is considered to enhance the local search capability. In addition, an efficient task scheduling using the Fuzzy Improved RFO (FIRFO) algorithm and game theory named EGFIRFO is presented considering four conflicting objectives (i.e., resource utilization, load balancing, makespan, and execution time). As a global optimization, the proposed method is compared with Bat algorithm (BA), Grey Wolf Optimizer (GWO), Particle Swarm Optimization (PSO), Antlion Optimizer (ALO), and Red Fox Optimizer (RFO) in terms of a set of qualitative parameters (i.e., convergence curve, trajectory, average fitness, and search history). The experimental results indicated that the proposed algorithm could improve the fitness value by 24.18%, 12.81%, 15.65%, 33.49%, and 14.73% compared with the BA, GWO, PSO, ALO, and RFO, respectively. As a scheduler, EGFIRFO is compared to three scheduling algorithms (i.e., MSDE, EMVO, and HGSWC).