The Web services model has emerged as a standard for representation, discovery, and invocation of services in a distributed environment. During service composition, Selection of appropriate services for composition is the chief task. A series of tasks are tied together to form a composite web service. Such composition is a challenging task because a number of Web services with the same or similar functions are increasing rapidly. QoS plays a vital role in the problem of selecting the most appropriate web service for composition. It is a measure for how well the composite web service serves the requester. Particle Swarm Optimization (PSO) algorithm and Gravitational Search Algorithm can be used to resolve this problem of optimal service selection. To compose individual web services dynamically according to users’ requirements in functional aspects as well as in non-functional preferences. In order to meet the QoS requirements of consumers, this paper presents the QoS calculation of non-functional requirements such as cost, availability, reliability and execution time. To verify the effectiveness in latency of Web Services selection the above two algorithms are compared. Results indicate the Gravitational Search Algorithm improves the latency over the Particle Swarm Optimization algorithm.