A neural network-based analysis method for the identification of a viscoplasticity model from spherical indentation data, developed in the first part of this work [J. Mater. Res.21, 664 (2006)], was applied for different metallic materials. Besides the comparison of typical parameters like Young’s modulus and yield stress with values from tensile experiments, the uncertainties in the identified material parameters representing modulus, hardening behavior, and viscosity were investigated in relation to different sources. Variations in the indentation position, tip radius, force application rate, and surface preparation were considered. The extensive experimental validation showed that the applied neural networks are very robust and show small variation coefficients, especially regarding the important parameters of Young’s modulus and yield stress. On the other hand, important requirements were quantified, which included a very good spherical indenter geometry and good surface preparation to obtain reliable results.