Sugarcane (Saccharum sp. hybrids) crops typically grow for 16–24 months in the subtropics, with nitrogen (N) fertiliser generally applied as a single dose between 150 and 250 kg N ha−1 early in the season. High N fertiliser application coupled with intense rainfall in the subtropics can lead to nitrate leaching and denitrification events that result in low N fertiliser use efficiency and damage to the environment. We investigated whether the use of a slow-release N fertiliser, polymer coated urea (PCU), may be more agronomically effective than urea as an N fertiliser source, by better matching soil N supply to sugarcane N demand. Multi-rate N fertiliser trials comparing biomass production and N accumulation responses of ratoon sugarcane crops to urea and PCU products were conducted across four commercial sugarcane farms in the Australian subtropics, with N fertiliser applied in a band 100–150 mm below the soil surface 2–12 weeks after the previous cane crop was harvested. At two sites, buried mesh bags containing 90 d and 270 d PCU products were destructively sampled over 12–15 months to assess the N release rate under field conditions. Sugarcane biomass yields were responsive to applied N at two of the four sites (P < 0.05) and crop N accumulation was responsive to N fertiliser application at all four sites (P < 0.1). While the mesh bag study clearly indicated a delayed release of N from the PCU products over time, there was no significant effect of N fertiliser source (urea vs PCU) on crop biomass or N accumulation at any site. The lack of any improvement in agronomic N efficiency with the PCU products is attributed to the presence of active roots in ratoon crops combined with the absence of large rainfall events in the months following N fertiliser application in the seasons of study. Modelling, coupled with an understanding of the N release dynamics of PCU products across a different soils and climatic conditions, is required to develop recommendations for PCU products for sugarcane growers in the region, although further trials across a wider range of seasons may be warranted to verify any modelling predictions.