A Surgical Planning System is proposed for Radio-Frequency Ablation (RFA) of liver tumors. The system includes a novel Stochastic Finite Element (FE) method which comprises the deterministic FE method with a probabilistic nature. The method enables modeling of variations in biological systems/processes which have properties that varied within a range of continuous values. By integrating Stochastic FE method in the simulation and surgical planning of RFA, complete tumor destruction could be achieved with minimum damage to healthy tissues. Stochastic simulation accounts for the variation in physical properties present in liver tissue which affects FE simulation accuracy as well as the effect of blood flow on RFA effectiveness due to the heat sink phenomenon. The proposed surgical planning method can be efficiently integrated into a surgical navigation system or robotic system.