The EarthCARE (EC) radar will be the first atmospheric cloud-profiling radar in space with Doppler capabilities. The mitigation of mispointing uncertainties related to thermal distortions and vibrations of the antenna structure, and/or attitude determination errors represents one of the several challenges to overcome for the reaching of high-quality Doppler velocity products. In addition to the possibility of adopting mispointing correction techniques based on the surface Doppler velocity in clear sky regions, we propose to use the presence of abundant natural targets (ice clouds) with climatologically very well characterized Doppler velocities as calibration points. This is demonstrated by showing that the ice clouds present on average in each EC orbit (emulated with CloudSat data) can be used to accurately retrieve the effect of an orbital harmonic, which represents the most relevant component in the modeling of the antenna mispointing error expected for the EC satellite. The residual errors associated to the orbital harmonic mispointing are negligible (< 0.03 m/s) compared with the other errors typically involved in Doppler velocities observed from fast-moving platforms. On the other hand, high-frequency components of the torque spectrum cannot be mitigated and, according to the current thermal and mechanical models for the EC antenna, are estimated to contribute to an RMS error of 0.2 (0.3) m/s in the nominal (worst case) scenario.