Colorimetric models are used to predict printer output colors from a set of toner control values. One such a model is the Neugebauer model, which predicts the spectral reflectance of the printed colors. This paper presents a new method for estimating the parameters (dot growth functions and primary reflectances) of the Neugebauer model from a set of measured spectral reflectances. Using information on the size of the uncertainty, a robust estimation algorithm (REA) is proposed to obtain a model with low sensitivity to undesirable and unpredictable variations and uncertainty. A case study using a high-end color printer shows that REA is consistently more robust than existing methods.