Exact diagnosis of Fracture Healing period is a challenging task to medical practitioners. Recent studies have addressed the problem of ‘fracture reunion prediction’ by different methods including electrical stimulation approaches. In this work, comparison of fracture healing period diagnosed using mathematical modeling of tibia fracture and Neural Network is presented. Using the electrical data recorded across 32 different tibia fracture patient’s empirical models like FOPDT (First order plus dead time), FOPDTZ (First order plus dead time and Zero) and higher order model were developed. The 32 patients were classified into 4 group’s namely fresh presentation, presentation after a medium delay, presentation after a long delay and facture with gap. Neural network was trained using electrical data recorded across different tibia fracture patients whose fracture site was stabilized using Teflon coated rings and a DC input voltage of 0.7V was applied via K-wires. A three layered feed forward neural network model designed using Levenberg-Marquett (LM) Back propagation training algorithm was able to predict the fracture reunion prediction with Relative Absolute Error (RAE) of 0.12 to 3.5. © Research India Publications.