The inherent difference between the modeled terrain and observed topography has seriously affected the 2 m temperature verification accuracy.The traditional 2D interpolation scheme can only ensure the consistency in two-dimensional space at latitude and longitude,while its neglect of the consistency at vertical direction produces error between forecast result and observation data.Using three-dimensional forecast variables,combined with near-surface elements of forecast products,an advanced 3D interpolation scheme is developed to ensure the consistency with the observed 3D space forecasting.The 24 h forecast products and observation data in July 2013 are used to compare and the results show that the proposed 3D interpolation scheme effectively solve the evaluation difference caused by the height bias between the modeled terrain and observed topography,which have maintained relatively consistent error distribution under different terrain height.And the evaluation results are not affected by terrain,which means it has good application value.To verify the proposed 3-D interpolation scheme for complex terrain with different resolutions,we carry out 60 h simulation sensitive tests for 2 m temperature estimation of Tibetan in July,2012 under resolution of 45 km,15 km,5 km,and 2 km,which show that the proposed method can improve estimation accuracy even under high resolution of 2 m.