Abstract: Doppler velocity information can be used for removal of ground clutter, extraction of wind profiles, detection of shear zones, and construction for dual Doppler wind fields. Operational application of Doppler velocity data from weather radars is hampered by the infamous limitation of the range velocity ambiguity. The dual pulse repetition frequency (DPRF) is a commonly used technique, especially in C band new generation weather radars that extend the unambiguous Doppler velocity. However, the region of chaos radial velocity that contained opposite sign velocity values with large differences often occurs in velocity data which obtained from C band new generation weather radars running in the dual PRF mode. It affected radar data quality severely. The characteristics of the chaos radial velocity are analyzed in this paper, and an algorithm for identifying and correcting the chaos radial velocity has been developed and examined based on the volume scanning radar data from Haerbin and Mudanjiang radars in the summer of 2010. The main conclusions are gotten as follows: (1) The signal to noise ratio has definite effects on local velocity change. It is found that where the signal to noise ratio is smaller, the local velocity change is bigger and the probability of chaos radial velocity occurred in velocity data is greater, and vise versa. (2) In the region of chaos radial velocity, the right velocity value is related to the unambiguous Doppler velocity value and the error velocity value. Subtracting the unambiguous Doppler velocity value from the error velocity value when the error velocity value is greater than zero, or adding the unambiguous Doppler velocity value to the error velocity value when the error velocity value is smaller than zero, then we could get the right velocity value. (3) The algorithm uses characteristics of velocity difference that in small region identify the region with big local velocity change, and turns out two characteristic parameters, to distinguish the chaos radial velocity with the margin zone of the velocity aliases kept well. That represents the abrupt changes in region, and the absData shows the average velocity differences in small region within the interval of 0 to 2 ×(Nyquist velocity). And, (4) on the basis of image continuity, the average velocity value with maximum ratio between the positive velocity points and the negative velocity points which are not the chaos radial velocity and are all in the region of K neighborhood can be acted as the new velocity value. In comparison with original velocity data, the region of chaos radial velocity has been retrieved obviously, and the quality of velocity data has been improved.