The virtual force method is applied for deploying a mobile sensor network in an uncertain environment. A mobile sensor network is a distributed collection of nodes, each of which has sensing, computation, communication and locomotion capabilities. Previous studies of mobile sensor deployment adopted a round by round process, that is, sensors move iteratively until the maximum coverage is reached. However, the mobile sensors probably move in a zig-zag way that waste a lot of energy in comparison with that of moving directly to a final location. We proposed a cluster-based virtual force algorithm (CVFA), which is an energy efficiency deployment strategy. CVFA can enhance the coverage after an initial random placement of sensors. By constructing virtual force, each node is attracted or repelled by other nodes, thereby forcing the network to spread itself throughout the environment. In CVFA, global information of k-hop neighbours can be acquired by the inter-cluster communications, and sensors can grasp their target locations logically. Experimental result, using ns-2 simulator, suggests that CVFA can significantly reduce the energy consumption compared to previous work, especially on maintaining similar coverage.