This paper explains an image-based force estimation of deformable tissue for a small single-port surgical robot. Since the end-effectors of a single-port surgical robot are thin and slender, it is one of challenging issues to install a force-torque sensor on its tip. Instead of using a force sensor, image-based estimation of interacting forces between deformable tissue and surgical end-effectors are proposed in this paper. For this purpose, a depth map extracted from stereo-endoscopic images is compared with an original shape of the deformable tissue which is obtained from the pre-operative medical images. To mitigate computational burden and to distinguish interaction forces of two end-effectors, subsets of whole nodes near contact points are utilized to calculate the interaction forces. This image-based force estimation will be integrated with the developed single-port neurosurgical robotic system to show its efficiency and performance.