Detecting inhomogeneity in a turbid tissue using an optical tomography technique as a non-invasive tool is of paramount importance in imaging and diagnosis of diseases and tissue abnormalities. In our approach, an ultrafast laser is used as the detecting source and the backscattered light signals are collected around the boundary of the target. We have developed a Monte Carlo program used to simulate time-dependent photon transport in inhomogeneous turbid media. As the laser pulse is attenuated by absorption and scattering, the detected temporal signal depends strongly on the optical properties of the medium and therefore leads us to the detection of inhomogeneity. Simulation results have shown that the presence of a small absorbing inhomogeneity in a highly scattering tissue will yield different log slopes in the temporal intensity profile. Further experimental studies in this paper yield similar results.