Most traditional force control methods for robotic system are based on the assumption that the interaction model is purely elastic. However, in the scenario of robotic-assisted surgery, it has been shown that the interaction between robotic instrument and the soft human tissue exhibits much more complex behavior involving change rate of contact position and force etc. In this work, an adaptive force tracking control algorithm has been developed based on a viscoelastic model (Kelvin-Boltzmann) and it can online adapt to interaction model parameter estimation errors. Physiological motion of tissue is also considered in the control design. The force tracking error is guaranteed to converge asymptotically even with parameter mismatches in the interaction model. Simulation studies were carried out to show performance improvement of the developed adaptive force tracking control algorithm over control method without adaptation to model parameter uncertainties.