Abstract The Weather Research and forecasting (WRF) ensemble three-dimensional variational data assimilation (3DVAR) hybrid system is applied to the assimilation of radial velocity (Vr) data from two WSR-88D radars for the prediction of Hurricane Ike (2008) before and during landfall. The existing hybrid system is upgraded by the vertical localization through Empirical Orthogonal Function decomposition. The hybrid ensemble member analysis is calculated through the cost function with perturbed observations. The Vr data are assimilated every 30 minutes for 3 hours immediately after Ike enters the coverage of the two radars. Subsequent 21-h Advanced Research WRF model deterministic forecast is performed to evaluate the data assimilation. The WRF 3DVAR system assimilates the same radar data for the comparison. The results from the hybrid method have shown that the wind and temperature analysis increments are consistent with observed hurricane circulations. The final analysis hurricane intensity and position closely follows the best track estimates. Meanwhile, the hurricane structure such as the eye, eye wall, and warm core is well simulated. All data assimilation experiments show improved intensity analysis and forecast compared to the experiment without data assimilation. Compared with 3DVAR forecast results, the 21-h forecast Hurricane Ike position is closer to the best track and the 6-h forecast root mean square error of Vr is relatively smaller. Additionally, the 3-h accumulated forecast precipitation equitable threshold score scores, with observation of National Centers for Environmental Prediction Stage-IV precipitation analysis, are much higher than 3DVAR forecast during the first 12 hour forecast. All results of the hybrid method are promising and future work to assimilate both Vr and reflectivity will be beneficial.