The capability of a previously proposed VDM (vehicle dynamic model) based navigation method for UAVs is assessed in attitude determination without IMU data. This method utilizes the VDM as main process model within the navigation filter and treats data from IMU (if available) and other sensors as observations. Experimental results from a test flight and a Monte Carlo simulation are presented to compare the performance of the VDM-based navigation against the conventional INS-based navigation. While the attitude estimation performance of VDM/GNSS integration is comparable to that of INS/GNSS in simulations, experimental results reveal the need for improved VDM parameter calibration.