In this paper, we propose a modified incremental subgradient (MIG) algorithm with a variable step size for positioning and tracking a target in wireless sensor networks. The proposed positioning scheme formulates location estimation as a nonlinear least-squares problem using the received signal strength, and then applies the MIG algorithm with a fixed step size to solve the problem. This scheme can be realized in an iterative, decentralized manner to improve both bandwidth and energy efficiencies. To track a moving target, we further present a step-size adjustment mechanism based on the velocity of the target. In addition, a convergence analysis is given for the MIG-based positioning process. As compared with related positioning and tracking methods, the proposed scheme has better location accuracy and tracking performance.