Presented at the 2016 Ocean Sciences MeetingSSS has been measured from space for the past 6 years with the SMOS and Aquarius missions. These two missions should have filled the gaps in the current in-situ network. Few data assimilation experiments have been realized. It is largely due to large errors and biases in the data. This needs to be addressed before assimilation in operational ocean forecasting systems. Our previous SSS data assimilation studies have shown that removing the systematic bias was a key issue. In this study, we propose to estimate and remove the large scale bias with the operational ocean forecasting system at 1/4°. The bias correction method is based on a 3D-Var method already used for correcting the model bias with in-situ data. Results show that unbiased SMOS has a positive impact. It helps to fill the gap in particular in the tropical convergence zones.