A new method is proposed for estimation of nonlinear elastic properties of soft tissues. The proposed approach involves a combination of nonlinear finite element methods with a genetic algorithm for estimating tissue stiffness profile. A multipoint scheme is introduced that satisfies the uniqueness condition, improves the estimation performance, and reduces the sensitivity to image noise. The utility of the proposed techniques is demonstrated using optical coherence tomography (OCT) images. The approach is, however, applicable to other imaging systems and modalities, as well, provided a reliable image registration scheme. The proposed algorithm is applied to realistic (2D) and idealized (3D) arterial plaque models, and proves promising for the estimation of intra-plaque distribution of nonlinear material properties.