Faced with increasing demand for network services by a huge number of users, requests to the web servers have significantly skyrocketed. Consequently, most of these servers need to run twenty-four hours a day, seven days a week with a high reliability and availability. Thus, the tremendous growth of the Internet has led the requirement of web server cluster management in order to deal with these issues effectively. Without using efficient mechanisms, an overloaded web server cannot provide great performance. In clusters, this overloaded condition can be avoided using load balancing mechanisms by sharing the load among available web servers. The existing load balancing mechanisms which were intended to handle static contents will be deprived from substantial performance under database-driven and dynamic contents. The most serviceable load balancing approaches to provide better results under specific conditions are Weighted Round Robin (WRR) and Client Aware Policy (CAP). By considering this, a Multi-Class Load Balancing algorithm (MCLB) was proposed for web server clusters and also an analytical model was proposed for calculating the load of a web server. The requests are classified based on the service time and keep tracking the number of outstanding requests at each webserver to achieve better performance. The service time of each request class type is used for load balancing. The experimental results demonstrate the effectiveness of the proposed approach by improving the average response time, error rate and throughput of the web server cluster. © 2019 Elsevier B.V., All rights reserved.