Sugarcane straw is a potential feedstock to increase bioenergy production. However, leaving straw on the soil surface is important to sustaining soil organic matter (SOM). We conducted a field experiment to evaluate the impacts of two years of straw removal on SOM in a sugarcane area – on an Eutric Nitisol in south-eastern Brazil – with low straw production (7 Mg ha–1). The experiment was a randomised complete block design (four blocks) with four rates of straw removal: (i) no removal, (ii) medium removal, (iii) high removal and (iv) total removal. We evaluated straw decomposition, soil carbon (C) and nitrogen (N) stock (0–30 cm), labile C, and microbial biomass C and N (0–10 cm). Increases in the straw removal rates reduced decomposition, with 30–42% of the initial straw amount remaining on the soil after one year (first year) and 33–38% remaining at the end of the second year. During the first year, no effect was found on soil C (8 kg m–2) and N (0.7 kg m–2) stocks; however, during the second year, total straw removal reduced soil C stock by 10% (0–30 cm) compared with no straw removal. Straw removal rates negatively affected the labile C (6–10% and 20–25%, first and second year respectively) and microbial C (correspondingly 10–15% and 21–27%). Microbial N was not altered in the first year, but decreased 26% due to straw maintenance (no removal) during the second year. Total straw removal decreased the C management index (CMI) by 15% (first year) compared with no straw removal. However, during the second year, all rates of straw removal decreased the CMI: by 15% for medium, 22% for high and 26% for total removal. Our findings indicate that any straw removal in areas with low straw production negatively affects SOM, initially through reduction in labile C pools. Although slight impacts on soil C and N stock were detected, reduction in straw-derived inputs stimulates microbial utilisation of soil C pools, and consequently reduces SOM over time.