Abstract: Pollination is one of the most important ecosystem services provided mainly by bees for crops and wild plants. The universality, adaptability and diversity of wild bees are driven more as reliable pollinators than honey makers. However, the diversity of wild bees had significantly reduced in recent years due to the loss of habitat and the erosion of habitat quality, which potentially threaten pollination service and agricultural sustainability supported by bees. The functional group diversity of wild bees is also the important component of diversity, which is a better predictor of pollination than species diversity. Wild bee species with different functional traits need different resources and respond to environmental changes differently. However, there is little knowledge on the response of different functional groups of wild bees to habitat types. To determine the spatio-temporal dynamics of wild bees in different habitats and the importance of different habitats for wild bee conservation in agricultural landscape mosaics, netting-catching was used to collect wild bees in four habitats (planted woodland, natural shrub forest, wasteland and peach orchard) during the period from April to September 2016. The temporal dynamics of diversity of wild bees and their functional groups in different habitats were analyzed. The results showed that natural shrub forest had the highest individual numbers and species richness of wild bees, large bees, solitary bees; and highest individual number of medium-sized bees. While planted woodland had more individuals of small bees, peach orchard had the lowest individuals number and species richness of wild bees. Natural shrub forest, planted woodland and peach orchard all had similar temporal dynamics of wild bee diversity, with one peak in spring and the other in summer. Overall, all functional groups of wild bees had the highest diversity in natural or semi-natural habitats across the sampling seasons. Peach orchard had the lowest wild bee diversity in May, which was attributed to intensive management of groundcover in peach orchards. Similarly, there were more unique species solely in natural shrub forest because of high resource availability and low human disturbance. Other habitat types also had some unique species which contributed to the overall bee diversity in the region. The four most abundant species were respectively Nomia thoracica, Halictus aerarius, Ceratina flavipes and Lasioglossum halictoides, which had different temporal dynamics in the habitat types. The species N. thoracica and C. flavipes were distributed in all the habitat types, except for wasteland. While N. thoracica was active in June through August, C. flavipes was active throughout all the sampling seasons. H. aerarius and L. halictoides existed in all habitat types, with the former active from May through September and the latter from April through July. Therefore, the four habitat types all had essential resources for wild bees, but the natural and semi-natural habitats were more important. In particular, natural shrub forest sustained more bees and had greater pollination efficiency. It was concluded that habitat diversity (especially natural and semi-natural habitat diversity) and less intensive management during active seasons of wild bees were critical for improving wild bee diversity and pollination services.