This study uses geostatistical modelling to identify hot spots of metal contamination produced by hazardous mining tailings from an inoperative Portuguese mine. Lousal is one of the many massive sulphide deposits of the Iberian Pyrite Belt. The mine is now closed, but the heavy metal enriched tailings remain at the site. Applying geostatistics to the diagnosis of this potentially contaminated area we aim to assess the quality of the soil, namely through a risk probability mapping for arsenic, for a better knowledge about the vulnerability of the soil to arsenic contamination. To achieve this aim, the initial variable was transformed into two indicator variables using the risk-based standards (intervention values) for soils, as proposed by Swartjes (1999), as the limit value. To account for the spatial structure, sample variograms were computed for the main directions of the sampling grid and a spherical model was fitted to the variables (arsenic raw data and indicator variables). The parameters of the model were used in the estimation process. To assess the vulnerability of the soil towards the mining works, a soil quality probability mapping for arsenic was carried out showing contamination probabilities in the area. The use of indicator kriging, as an alternative to other more complex kriging methods, in the soil data of the Lousal mine, produced unbiased soil quality maps.