Cloud Computing makes use of the Internet to deliver standard and virtualized resources, software, and data as a service. Users will have the option to take use of all of these advancements in information technologies with cloud computing approach without needing to be extremely informed or competent in each one separately. An enormous number of services are currently available. One of the main challenges in the cloud computing environment is the lack of a workable method for developing online web services that may achieve high potential for Quality of Service while following to service level agreement (SLA) boundaries. Choosing or composing those combined services is referred to as nondeterministic polynomial time optimization problem. Many metaheuristic algorithms have been used in the past due to the NP-hard complexity of service composition. The reliability and QoS (Quality of Service) issues with other grid computing models are addressed. This paper presents the Coaching Based Multi-Verse Optimization (CBMVO) methodology in Web services, covering their composition and implementation methods as a metaheuristic algorithm to achieve the desired goals. The application of optimization techniques to choose web services in cloud settings with a normalised quality of service (QoS). By considering various QoS metrics, employing a hybrid optimization approach, and taking user preferences into account throughout the selection process, the method outperforms previous methods. These developments lead to a more effective and customized choice of web services, eventually enhancing user experience across cloud environments. The proposed methodology is evaluated and the results are validated the efficiency and performance of the proposed approach in relation to reliability, availability and cost. The new algorithm for quality of service stands out from other methods because it has a higher score of 0.89 while the MFO method only scored 0.77. This means that the suggested technique has improved the quality of service by 13% compared to the CBMVO method.