Correlation analyses have been conducted between pollen assemblages of Tauber traps and surface soil samples in the steppe area of northern China. The results indicate that the pollen assemblages from these two sampling methods have a good similarity， ie. arbors content is lower than 10％〖KG-*5〗， shrub content lower than 5％ and herbs content higher than 75％. The pollen types held by all assemblages constitute 743％ of the total amount. Artemisia and Chenopodiaceae are the main components of pollen assemblages. Ephedra and Nitraia usually occur in the desert steppe areas， while Stellera and Cyperaceae pollens dominate the typical steppe areas. The DCCA analyses show that pollen assemblages have the most obvious relationship with the severe cold month temperature and a more obvious relationship with the annual precipitation. The pollen components of Tauber traps are different from those of the surface soil samples. The former has more pollen types than the latter; the pollen content of Artemisia and Chenopodiaceae is lower than 70％〖KG-*5〗， while the latter higher than 70％. The pollen content of the desert steppe areas is lower than that of the typical steppe areas， which indicates vegetations are well developed in the typical steppe areas and their pollen production higher than the desert steppe areas， the pollen density of the surface soil samples in typical steppe area is obviously higher than in desert steppe areas which is due to the well developed vegetation and higher pollen production of the desert steppe area， and the poor preservation of the pollens in the desert steppe area. The Tauber traps samples and surface soil samples both can distinguish desert steppe area from typical steppe area， while cross experiment results of Tauber traps sample are better than surface soil samples.