Abstract Nitrogen recommendation systems for sugarcane (Saccharum spp.) generally does not consider the N supply from soil. Identifying a reliable soil test for estimating N availability is crucial to avoid yield losses or environmental pollution. Therefore, the objective of this study was to correlate and calibrate N availability indices with field–based measures of soil N supply. Between 2006 and 2013, 15 trials for rate–response to N fertilizer by sugarcane ratoons were performed in Sao Paulo, the main sugarcane–producing state in Brazil. The following indices were tested: KMnO4 oxidizable C, hot KCl extractable N, phosphate–borate buffer distillable N, NaOH distillable N, Illinois Soil N Test, organic C, total N, mineral N, anaerobic incubation, soil respiration, substrate–induced respiration, microbial biomass C, metabolic quotient, microbial quotient, and gross N mineralization. The indices were then correlated with sugarcane yield (Y0N) and N content of the crop (N0N) in N–unfertilized plots, relative yield (RY), and the N rate predicted to achieve 90% of the RY (NR 90%RY). Although weak correlations were found between Y0N with anaerobic incubation, total N, and soil respiration, as well as between N0N and anaerobic incubation, no index correlated with RY or NR 90%RY. Grouping sites based on soil texture or byproduct management did not improve prediction of RY. Therefore, we concluded that none of the fifteen laboratory indices is a reliable predictor of soil N supply, and hence could not be used to adjust N fertilization rate for sugarcane.