To conserve and manage the bat population in an agricultural landscape, we investigated the relative importance of environmental factors on their occurrence and abundance in riparian zones at site and landscape scales. From May to October 2003, four bat species (Myotis daubentonii, M. frater, Plecotus auritus, and Eptesicus nilssonii) were mist-netted at 26 sampling sites in five streams in the Tokachi plain in central Hokkaido, northern Japan. The environmental factors of the two scales were measured at each sampling site. To clarify the most suitable landscape scale for bats, we also compared three spatial extents (250, 500, and 750 m) at each site. Generalized linear mixed models revealed that the occurrence and/or abundance of all four species, except for M. frater, was positively correlated with the percent cover of broadleaved deciduous forest and the height of bank-side tree canopy. Additionally, according to variation partitioning, although the value of the pure effect varied among species and measurement types, the site scale had a relatively larger pure effect than the landscape scale in many cases. However, in three of the five combined models, the proportion of confounding effect was much larger than that of the pure effect at the respective scales. Our study demonstrated that streams adjacent to a well-developed broadleaved deciduous forest are important for night habitats of many bat species in an agricultural landscape. Although focusing on the site scale may be important for bat conservation, we strongly stress the necessity for conservation and management plans at multiple scales.