Over the past three decades in boreal Sweden, there has been a long-term decline of cyclic sympatric voles, leading to local extinctions of the most affected species, the grey-sided vole (Clethrionomys rufocanus). We monitored this decline by snap-trapping on 58 permanent plots spread over 100 km2 in spring and fall from fall 1971-2003. The reason for the decline is largely unknown, although a common major factor is likely to be involved in the decline of C. rufocanus and of the coexisting voles. However, here we deal with the reasonability of one complementary hypothesis, the habitat fragmentation hypothesis, which assumes that part of the decline of C. rufocanus is caused by habitat (forest) destruction. There was considerable local variation in the decline among the 58 1-ha sampling plots, with respect to both density and timing of the decline; however, all declines ended up with local extinction almost without exception. Local declines were not associated with habitat destruction by clear-cutting within sampling-plots, as declines started about equally often before as after clear-cutting, which suggested that habitat destruction outside sampling plots could be involved. In a multiple regression analysis, local habitat preference (LHP; expressed as a ratio of observed to expected number of voles trapped per habitat) together with two habitat variables in the surrounding (2.5×2.5 km2) landscape matrix explained 56% of the variation among local cumulated densities of C. rufocanus and hence of local time-series. LHP was positively correlated and explained 31% of the variation, while connectivity among clear-cuts was negatively correlated and proximity among xeric-mesic mires was positively correlated and explained additional 16% and 9%, respectively. Even if the overall decline cannot be connected to local clear-cutting on sampling-plots, clear-cutting and hence habitat fragmentation/destruction in the surrounding landscapes potentially influenced grey-sided vole numbers negatively.