Wind turbine blades often present complex distributions of their stiffness, mass and inertial properties along the span. We propose a method to estimate such physical parameters so as to match given experimental observations. The procedure can be used to understand the nature of possible discrepancies between designed and manufactured blades and to provide updated high fidelity mathematical beam models to be used in aero-elastic simulations. The formulation is based on the constrained optimization of a maximum likelihood cost function and a noisy measurement fusion approach whereby the data of multiple experiments are used simultaneously in a single estimation process. The proposed method is demonstrated first using simulated data and then in the identification of two real small wind turbine blades. Copyright © 2012 John Wiley & Sons, Ltd.