Axial resolution and traverse resolution in optical coherence tomography (OCT) imaging are determined by different factors, while axial resolution is determined by both the coherence length of light source and the beam-focusing condition, and traverse resolution is determined by the beam-focusing condition of the sample arm. In the main approaches to axial resolution improvement in OCT, a light source with a broaden bandwidth is used and coherence gating is combined with apodization, which cannot improve the traverse resolution. A method is introduced to increase both the axial resolution and traverse resolution simultaneously in an OCT system by the path length code and coherent compounding method. Different effective functions are formed by adding a path length coding lens in to the proposed OCT system, which are corresponding to different path lengths. Owing to the intrinsic ability to differentiate path lengths, we can obtain several images of the same sample, corresponding to the different effective functions simultaneously. By adding these functions through numerically controlling their relative contributions, we can finally obtain a coherent compounding signal with three-dimensional superresolutions of axial resolution and traverse resolution. Compared with the previous approaches, the path length code and coherent compounding method is very easy to operate and its cost is very low, which can not only avoid the high cost and inconvenience in implantation, but also increase both axial and traverse resolutions simultaneously.