Hyperspectral infrared (IR) sounders provide high vertical resolution atmospheric sounding information that can improve the forecast skill in numerical weather prediction (NWP). Commonly only clear radiances are assimilated, because IR sounder observations are highly affected by clouds. A cloud-clearing (CC) technique, which removes the cloud effects from an IR cloudy FOV and derives the cloud-cleared radiances (CCRs) or clear sky equivalent radiances, can be an alternative yet effective way to take advantage of the thermodynamic information from cloudy skies in data assimilation. This study develops a VIIRS based CC method for deriving CrIS CCRs under partially cloudy conditions. Due to the lack of absorption bands on VIIRS, two important quality control steps are implemented in the CC process. Validation using VIIRS clear radiances indicates that the CC method can effectively obtain the CrIS CCRs for FOVs with partial cloud cover. To compare the impacts from assimilation of CrIS original radiances and CCRs, three experiments are carried out on two storm cases, Hurricane Joaquin (2015) and Hurricane Matthew (2016), using GSI assimilation system and WRF-ARW models. At the analysis time, more CrIS observations are assimilated when using CrIS CCRs than with CrIS original radiances. Comparing temperature, specific humidity, and U/V winds with radiosondes indicates that the data impacts are growing larger with longer time forecasts (beyond 72-hour forecast). Hurricane track forecasts also show improvements from the assimilation of CrIS CCRs due to better weather system forecasts. The impacts of CCRs on intensity are basically neutral with mixed positive and negative results.