This study presents an approach to assimilate tropical cyclone (TC) real-time reports and the University of Wisconsin-Cooperative Institute for Meteorological Satellite Studies (CIMSS) Atmospheric Motion Vectors (AMV) data into the Weather Research and Forecasting (WRF) model for TC forecast applications. Unlike current methods in which TC real-time reports are used to either generate a bogus vortex or spin up a model initial vortex, the proposed approach ingests the TC real-time reports through blending a dynamically consistent synthetic vortex structure with the CIMSS-AMV data. The blended dataset is then assimilated into the WRF initial condition, using the local ensemble transform Kalman filter (LETKF) algorithm. Retrospective experiments for a number of TC cases in the northwestern Pacific basin during 2013–2014 demonstrate that this approach could effectively increase both the TC circulation and enhance the large-scale environment that the TCs are embedded in. Further evaluation of track and intensity forecast errors shows that track forecasts benefit more from improvement in the large-scale flow at 4–5-day lead times, whereas the intensity improvement is minimal. While the difference between the track and intensity improvement could be due to a specific model configuration, this result appears to be consistent with the recent reports of insignificant impacts of inner core data assimilation in operational TC models at the long range of 4–5 days. The new approach will be most beneficial for future regional TC models that are directly initialized from very high-resolution global models whose storm initial locations are sufficiently accurate at the initial analysis that there is no need to carry out any artificial vortex removal or filtering steps.