This paper presents a novel approach for grasp optimization considering contact position and object information uncertainties. In practice, it is hard to grasp an object at the designated or planned contact positions, as errors in measurement, estimation, and control usually exist. Therefore, we first formulate the influences of contact uncertainties on joint torques, contact wrenches, and frictional condition. We then include external wrench uncertainties in the required external wrenches set. Based on this formulation, we define the linear grasp optimization problem for two kinds of frictional contact models-frictional point contact and soft finger contact-so that we can successfully grasp an object even if deviations in contact point, object weight, and center of mass occur. The validity of our approach is shown by means of numerical examples and the result of experiments.