A realistic estimate of N mineralized from soil organic matter is essential for determining the rate of N fertilizer application required to optimize crop yield and quality and minimize adverse impacts of excessive N on the environment. We compared several indices (soil properties and chemical methods) against net-N mineralization determined by anaerobic incubation (AIN) to predict N mineralization over a range of soils under sugarcane. Soil sampling (0-20 and 20-40 cm) was carried out in 21 N-response sugarcane trials in São Paulo State, Brazil. The AIN was performed during 7 d at 40°C. Soil properties evaluated to estimate soil N mineralization were pH (0.01 mol L^-1 CaCl₂), sand, silt, clay, soil organic C (SOC) and soil total N (STN). Chemical methods assessed were phosphate-borate buffer at pH 11.2 (PBB), direct steam distillation with NaOH (DSD) and hot KCl-extractable N (HKCl). The PBB and HKCl indices were determined by counting the preexisting soil NH₄⁺-N. All indices were significant correlated with AIN. The SOC, STN, PBB, DSD and HKCl were moderately correlated to AIN (r ranging from 0.60 to 0.69) and had linear regressions with r² ranging from 0.36 to 0.47. The pH, sand, silt and clay provided poor correlations with AIN (r < 0.58; r² < 0.33). The pH, sand, SOC, STN, PBB and DSD produced the best linear multiple regression by stepwise procedure for predicting AIN (r² = 0.79). For this study, no single N availability index has proven robust enough for broad acceptance. However, our results suggest that a combination of laboratory methods can be useful for predicting net-N mineralization for sugarcane cropped soils.