Increased knee flexion at initial contact (KFIC) is a common gait deviation in Cerebral Palsy. Hamstring lengthening surgery (HL) aims to decrease KFIC, but may affect pelvic tilt (PTIC). In this work, we design a decision-making tool that simulates the effect of orthopedic surgery, with or without HL, on KFIC and PTIC. The postoperative parameters are estimated given preoperative gait, physical examination and surgery. Nonlinear regressions are performed by feedforward neural networks. On test, the mean root-mean squared error for KFIC and PTIC are 9° and 5° respectively. Sixty-three percent and 90% of the lower limbs are estimated within a 10° error range for KFIC and PTIC respectively. The simulator is able to give good estimations independently of the surgery type.