Thermal anemometers with omnidirectional sensors are recommended in the standards to be used for low-velocity measurements indoors. Requirements for the directional sensitivity of the velocity sensor are prescribed. However, a method for testing the directional sensitivity of low-velocity anemometers does not exist. A simple test procedure is used in practice to identify the so-called "yaw" and "roll" directional sensitivity of an omnidirectional velocity sensor. However, this procedure cannot be used to assess the way in which the accuracy of the velocity measurements is influenced in practice by the directional sensitivity of the sensor. In this paper, a test method for describing the directional sensitivity of omnidirectional velocity sensors is proposed and used experimentally. The test method can be used to define the impact of directional sensitivity on the accuracy of the velocity measurements. Further, the proposed method can be used to improve the accuracy of the velocity measurements by optimizing the static calibration of low-velocity anemometers and the positioning of the velocity probe during field measurements. The method is suggested for inclusion in future indoor climate standards.