Numerous empirical and analytical relations exist between shield tunnel characteristics and surface deformation. Artificial neural networks (ANN) was used to develop predictive relations between the maximum surface settlement and shield tunnel overburdens, shield diameters, thrusts of shield tunneling, advancement rates of shield, fill factors of grouting, cohesive forces, friction angles and compression modules of the soils. So, ANN can become a useful predictive method. With the advantage of ANN in nonlinear problem, the theoretical model to predict the maximum surface settlement is established. The agreement of the measured results with the actual situation of being predicted shows that the proposed model is satisfactory.