Many research topics in recent years have focused on clustering algorithms as a scalable routing protocol in wireless sensor networks. In these algorithms, by the use of data aggregation techniques and reduction in sensors that should make long distance communications, energy consumption in total network lifetime is decreased. One of algorithms proposed in this field is DECSA. Aiming at the problem of sensors limited energy, this algorithm is an improvement based on the classic clustering algorithm LEACH. By taking into consideration nodes distance from each other and residual energy of nodes, DECSA algorithm tries to select proper nodes as CH. Also, in this algorithm by the aim to produce a multi-level clustering hierarchy and based on CH's distance to base station, some CHs are selected as base station CHs. In this paper, based on DECSA, a load balanced clustering algorithm is proposed. This algorithm incorporates two thresholds in DECSA in terms of supporting minimum and maximum acceptable load. Thus, when cluster load in one frame is smaller than the minimum threshold, cluster's frame is expanded and when it is bigger than maximum threshold, a fraction of cluster's load in each frame is transferred to a 2nd CH node.