To elucidate the expansion of bamboo (Phyllostachys pubescens Mazel ex Houzeau de Lehaie) forests, we used multiple linear regression analysis and determined whether there were site differences for data obtained in Hirasawa, Otaki-machi, Chiba Prefecture, and Kofuki, Takehara City, Hiroshima Prefecture, Japan. Vegetation maps from 1984 and 2001 for Hirasawa, and from 1986Hirasawa, and from , 1996Hirasawa, and from , 2000Hirasawa, and from , and 2006 for Kofuki, were compared, and the annual expansion rate of each P. pubescens forest was calculated. We evaluated nine indices, including original bamboo forest area, neighbouring vegetation, slope inclination and aspect, distance from roads, and shipment of bamboo shoots. Shipment of bamboo shoots was a positive factor for P. pubescens forest expansion, whereas the proportion of adjacent short vegetation, northness (the cosine of slope aspect), and area negatively affected the expansion in Hirasawa (R 2 = 0.683). On the other hand, distance from roads and eastness (the sine of slope aspect) positively affected expansion, while slope inclination was a negative factor in Kofuki (R 2 = 0.942). We calculated that P. pubescens forests expanded at about 2% per year in regions of reduced shipments of bamboo-shoot harvest. We showed that not only natural conditions but also management factors affect P. pubescens forest expansion. Regional differences in natural and anthropogenic effects on expansion were also found.