Development of neuronal prosthetics, where neuronal activity is used to control artificial limbs, has so far relied on decoding kinematic parameters of movements, such as movement position or velocity. In addition to kinematic control, proper control of forces exerted by the prosthetic device is necessary for successful interaction with the environment. In our study, we analysed the possibility of classifying and decoding different grasp related forces during active grasping. Two macaque monkeys were trained to reach, grasp and pull an object in response to visual cues. Cues instructed the monkeys to grasp the object with one out of two grip types (precision or side grip) and pull the object with one of two different forces (0.5N or 2N). Monkeys obtained a food reward after successfully performing the instructed grip and pull. During the task execution, we recorded electrophysiological signals from the multielectrode arrays implanted intracortically in the hand and arm area of the monkey’s motor cortex. Six different parameters of the grip: four pressure forces on each side of the object, pull force on the object and the object displacement, were recorded simultaneously with the neuronal activity. Recorded neuronal activity was used to classify different grip types or loading forces, and to decode the continuous traces of different forces during the grip. Our results show that kinetic grip parameters can be decoded with high accuracy, thereby improving the feasibility of constructing fully functional anthropomorphic neuronal prosthesis that relies on kinetic (force) control.