Near infra-red optical computerized tomographic imaging through a turbid tissue by using time-resolved measurement has been gaining widespread attentions from the field of biomedical research for its non-invasiveness and ability of disclosing the functional information of tissue. A variety of techniques, aimed at implementing a clinically-useful system, have been proposed for eliminating the obstacles arising from multiple light scattering of biological tissue. Among these scheme, one of the primary goals has been to seek for a fast, effective and mathematically rigorous algorithm for the reconstruction tasks. In this paper we first present the basic principle of time-resolved optical tomography. The diffusion approximation-based photon transport model in a highly scattering tissue, which offers an advantage in computational speed in comparison with other stochastic models, and the procedure for solving this forward model by using the finite-element method are then accessed. Theoretically, an iterative Newton-Raphson algorithm for solving the inverse problem is introduced based on the implicit calculation of Jacobian of the forward operator. Numerically simulated images of absorbers embedded in a homogeneous tissue sample are reconstructed from either mean-time-of-flight or integrated intensity data for the verification of the approach. Finally, the experiment setup which are under construction in our Lab is described in detail.