The proposed NASA Earth System Science Pathfinder – Hydrosphere States (HYDROS) – Mission will use both L-band microwave coarse resolution radiometer and fine-resolution radar to make the first space borne observations of global soil water availability that enables new scientific investigations of atmospheric predictability and global change processes. As one of the efforts of HYDROS science team, we attempt to explore the feasibility for generating a median resolution soil moisture product by combining both radiometer and radar observations via a Kalman Filter data assimilation approach in addition to other optimal estimation approaches. Based on the simulated land surface soil moisture and the corresponding data of land cover type, soil type, land surface temperature and atmospheric forcing variables of the Red-Arkansas River Basin at 1km resolution, HYDROS science team created a simulated data set of 1.4GHz brightness temperatures at 36km spatial resolution with horizontal and vertical polarizations and 1.4GHz radar backscatters at 3km spatial resolution with HH, VV and HV polarizations using selected microwave emission and radar backscatter models. To retrieve surface soil moisture from these simulated HYDROS observations with instrument errors and input parameter uncertainties added, several Extended Kalman Filter (EKF) data assimilation algorithms were designed and tested again the soil moisture observations aggregated from the 1km soil moisture simulations. This paper will present the designs of these EKF data assimilation algorithms, results of the evaluations and discussions on the application of the data assimilation methods in soil moisture retrieving.