53 pollen traps and surface samples were collected in order to detect the characteristics of pollen assemblages and their relationships with vegetation and climate in 16 forest communities located in 10 mountains in northern China. The results show that 72% of the pollen taxa (80 taxa) are the same between the traps and the surface samples. The dominant taxa in the plant communities are consistent with the main pollen taxa in the pollen assemblages at the same sites. In Pinus plant communities, both Pinus pollen influx and concentration are higher, indicating the high pollen productivity and good pollen preservation ability of Pinus. In Picea and Abies plant communities, Picea and Abies have lower pollen influxes but higher concentrations, suggesting their low pollen productivities but better pollen preservation abilities. In Betula and Quercus plant communities, Betula and Quercus have higher pollen influxes but lower concentrations, revealing their high pollen productivities but poor pollen preservation abilities. The study of relationships between pollen and vegetation with discriminant analysis shows that pollen assemblages from both trap and surface samples can reflect the characteristics of different communities and distinguish different ecological areas, but surface samples can reflect the dominant components of communities much better than the traps. The study on correlations between pollen assemblages and climate with DCCA reveals that significant correlations exist between pollen assemblages and mean temperature of the coldest month (r = 0.84 for trap samples, r = 0.72 for surface samples), and then annual mean precipitation (r = 0.73 for trap samples, r = 0.71 for surface samples).