Diffuse optical tomography is to find the value of optical coefficients in a tissue using near infra-red lights, which is usually modelled by an optimization problem being composed of two steps: the forward solver to compute the photon density function and the inverse solver to update the coefficients based on the forward solver. Since the resulting problem is mathematically nonlinear ill-posed inverse problem and numerically large-scale computational problem demanding high quality image, it is highly desirable to reduce the amount of computation needed. In this paper, domain decomposition method is adopted to decrease the computation complexity of the problem. Among many methods of domain decomposition techniques, two level multiplicative overlapping domain decomposition method and two level space decomposition method are used to the forward and inverse solver, respectively. The convergence and computational cost of each method are described. And the efficiency of using combined two methods is verified by the implementation of reconstructing the absorption coefficient on square domain and thin domain.