Cloud computing is an emerging technology in internet world. Load balancing is helped to distribute the dynamic workload between multiple nodes to ensure that no single node is overloaded. The proper utilization of resources is achieved by load balancing. Resource intensity aware load balancing (RIAL) method was proposed for load balancing in cloud computing. The resources like CPU, bandwidth, and memory and storage space in physical machines (PMs) are virtualized in the form of virtual machines (VMs) in Cloud computing. The resources in PMs are consumed by each VMs. The various resources in each PM were assigned different weight values based on the resource usage intensity of the PM by RIAL method. Based on the weight values, the VMs were selected from heavily loaded PMs to migrate out other lightly loaded PMs during load balancing operation. The destination PMs were selected to migrate selected VMs by using multi-criteria decision making (MCDM) method in which only lightly loaded PMs are considered. In PM, some resources are over utilized while other resources are underutilized. So, it is possible that the heavily loaded PM nearer to PM of selected migration VM might have required resources to balance the load. So in this paper, the VMs which are selected for migration are mapped with destination PMs globally. It is achieved by considering both the lightly loaded and heavily loaded PMs as destination PMs. For global mapping process, the expected completion time of each job in VMs of heavily loaded and lightly loaded PMs are calculated which decides the destination PMs through MCDM method. Hence, load balancing in cloud is further enhanced by improved RIAL (IRIAL) method. The simulation result shows that the proposed IRIAL method has better computational cost, performance degradation and execution time for load balancing in cloud.