We describe a method for retrieving winds from colocated Ku- and C-band ocean wind scatterometers. The method utilizes an artificial neural network technique to optimize the weighting of the information from the two frequencies and to use the extra degrees of freedom to account for rain contamination in the measurements. A high-fidelity scatterometer simulation is used to evaluate the efficacy of the technique for retrieving hurricane force winds in the presence of heavy precipitation. Realistic hurricane wind and precipitation fields were simulated for three Atlantic hurricanes, Katrina and Rita in 2005 and Helene in 2006, using the Weather Research and Forecasting model. These fields were then input into a radar simulation previously used to evaluate the Extreme Ocean Vector Wind Mission dual-frequency scatterometer mission concept. The simulation produced high-resolution dual-frequency normalized radar cross-section (NRCS) measurements. The simulated NRCS measurements were binned into 5 x 5 km wind cells. Wind speeds in each cell were estimated using an artificial neural network technique. The method was shown to retrieve accurate winds up to 50 m/s even in intense rain.