With the ever growing server density and infrastructures encompassed in data centers, heat generation and power dissipation has increased drastically. This has brought a serious impact on system performance and reliability. Therefore, it has demanded an efficient thermal management, where thermal balancing is one of the various techniques. This thesis presents a design and implementation of an autonomous virtual machine (VM) migration in order to attain thermal balance in server cluster hosting the VMs. Two algorithms, based on server temperature readings, have been implemented to attain thermal balance in server cluster. The first algorithm employs temperature readings of all the servers in the cluster, to make autonomous VM migration by choosing the coolest server. Whereas, the second algorithm requires only the temperature reading of the server where the VMs are hosted on and the VMs learn their environment by trial and error to do the autonomous migration. The results showed that both algorithms were able to maintain thermal balance in the server cluster by applying the autonomous migration.