The Norwegian Climate Prediction Model (NorCPM) is developed at the Bjerknes Center for Climate Research (BCCR) in Norway. NorCPM combines the Norwegian Earth System Model (NorESM) and the EnKF for the purpose of climate predictions and long-term climate reanalyses. Recently, we have revised the scheme of sea ice concentration assimilation in NorCPM. Compared to the previous scheme, there is no longer degradation in ocean heat content at the middle latitude. It is likely due to the fact that the implementation of the assimilation of sea ice concentration in the previous scheme strongly reduces the influence of SST data near ice-covered regions and thus influences ocean circulation. We noticed that the new regional sea ice prediction skill verified over 2002-2020 is higher than the old one verified over 1985-2010. It can be caused by the revised scheme, modulation of variability or improved accuracy of other datasets (e.g., ARGO floats). We implemented the assimilation of sea ice thickness in NorCPM. The error of ice thickness is constrained to about 1 m during the ESA CCI period and 0.6 m during the C2SMOS period. We will also present the benefit of ice thickness assimilation on seasonal sea ice prediction.