The sustainability of land management was assessed for a sugar cane plantation using soil chemical and physical properties as indicators. The plantation (6000 ha) was established in 1979 on a broad alluvial plain and the majority of the soils are classified as Eutric and Mollic Fluvisols (73%) and Eutric Vertisols (23%). Average annual rainfall is about 2000 mm with a dry season from May to November. The sugar cane is not irrigated and receives only N fertilizer (90 kg N ha −1 yr −1). Detailed soil maps were used to make a selection (186 samples) of the existing soil chemical database (1979–1994) and for selecting sample sites in 1996. The pH of the topsoils had decreased from about 6.5 to 5.8 in both Fluvisols and Vertisols accompanied by a decrease in CEC and exchangeable K. Between 1979 and 1996, organic C contents declined from about 5.5 to 3.2 g kg −1. A significant decline in available P was found in Fluvisols (40 to 32 mg kg −1) and Vertisols (37 to 25 mg kg −1). Significant changes in soil chemical properties were mostly confined to the topsoil and differences between Fluvisols and Vertisols were relatively small. Changes in soil physical properties were assessed by measuring bulk density and infiltration under sugar cane and adjoining natural grassland areas. Bulk density and water intake were similar under natural grassland and within the sugar cane rows. The interrow had a significantly higher bulk density due to wheel traffic, which caused very low water intake. Bulk densities at which infiltration rates were severely reduced were slightly higher in Fluvisols than in Vertisols but for both Major Soil Groupings an increase of only 0.2 Mg m −3 was critical. Changes in soil chemical and physical properties indicated that land management is not sustaining the resource base for sugar cane cultivation in the long-term. Threshold values in soil chemical properties were not reached and they were in 1996 still favourable for sugar cane cultivation. The soil compaction, however, directly affect the sugar cane as it seriously reduce rooting. It is concluded that routine soil analytical data combined with data that can be easily collected suffices to make a general assessment of sustainable land management at a plantation scale.