Differences in soil carbon accumulation rates can markedly affect the sustainability of ecosystems managed for food and fuel production. We examined C accumulation and persistence in candidate biofuel cropping systems that differed in perenniality (annual vs. perennial) and biodiversity (monoculture vs. polyculture), five years post-establishment. In eight replicated systems at each of the two sites we measured active, slow, and resistance C pools via long-term incubations and acid hydrolysis. Cropping systems included four annual systems; no-till continuous corn, and each phase of a corn-soy-canola rotation, three monoculture perennial systems (swithchgrass, miscanthus, and hybrid poplars), and three diverse perennial systems (native grasses, early successional, and restored prairie) at a moderate fertility site in southwestern, Michigan (Kellogg Biological Station) and a high fertility site in south central Wisconsin (Arlington). At KBS we found substantial differences in active pools between annual and perennial polyculture crops. Active C under polycultures were over 2.5 times greater than continuous corn. Active C rates in the continuous corn were 236.8 ug C g-1 soil compared to 656.4 ug C g-1 soil, 500 ug C g-1 soil, 638 ug C g-1 soil in the native grasses, early successional, and the restored prairie, respectively. Amongst the monoculture perennials, it was only the poplar system that had significantly greater active C compared to continuous corn with an active C pool size of 695.7 (ug C g-1 soil). Ecosystem differences in the slow C pool were less apparent, and there were no differences in the resistant pool. At the more fertile Arlington site, the restored prairie system had significantly greater active C compared to all systems except the corn in rotation system (p<0.05), but ecosystem differences between annual and perennial systems were not as apparent. Total accumulation during the 322 day incubation period did not differ by site, however, proportional to total C pools, more C was stored in the active and slow C pool at KBS, while the majority of C at Arlington accumulated in the resistant pool. Patterns of particulate organic matter did not match long-term incubation results, where large site differences were visible (p<0.05). These results demonstrate that diverse perennial systems are effective at increasing C in the active pool, especially in less fertile soils and with time could lead to greater stabilization of C in more recalcitrant pools.