Abstarct: Due to the presence of several small to large chromite mines and its favorable geological makeup, the Sabzevar Ophiolitic Belt is one of the best endowed chromite belts in Iran. This study used and adapted mineral prospectivity mapping technology to better define and understand the spatial distribution of podiform chromite mineralization at the regional and local scales. The first step in achieving these aims was to translate all features related to chromite mineralization into separate data laxyers and combine these in the geographical information systems environment using a variety of techniques. Since random and systematic errors resulting from expert judgment and exploration bias cause uncertainty and low reliability of prospectivity models, this study adopted continuous and improved data-driven methods to identify exploration targets. To this end, regional and local exploration criteria were identified baxsed on a mineral systems analysis of the podiform-type chromite deposits as well as a review of the available exploration data. Next, the predictive patterns related to chromite mineralization were elicited and weighted continuously. At the regional scale, receiver operating characteristics curves and prediction-area plots were employed to measure the degree of spatial association between continuously-weighted evidence maps and chromite mineralization. The results indicated that the evidential maps used in this study are significant predictors and play an important role in the generation of the overall mineral prospectivity model. In addition, the predictive patterns related to chromite mineralization were weighted discretely, for comparison with continuously-weighted predictive patterns, and evaluated after integration of the laxyers. Data-driven neuro-fuzzy network and random forest methods as well as mathematical functions were utilized to integrate continuously-weighted evidence maps at the regional scale. Moreover, discretely-weighted spatial evidence values were also integrated using a neuro-fuzzy network approach. The results of this study demonstrate that the performance of the neuro-fuzzy targeting model, generated with continuously weighted spatial evidence values, is superior to that of the neuro-fuzzy model, generated with discretely weighted exploration evidence data. Amongst the regional scale prospectivity models, the random forest model performed best and, thus, was deemed most appropriate for selecting a chromite district for detailed prospectivity mapping at the local scale. Last but not least, an improved prediction-area plot was developed for evaluating the performance of spatial evidence maps and prospectivity models. The advantage of this new performance evaluation tool over existing methods is its capability to simultaneously consider three important input criteria, including (1) prediction rate of mineral deposit locations, (2) prediction rate of non-deposit locations, and (3) area occupied by exploration targets in the evaluation scheme. In this study, fuzzy gamma and geometric average operators were utilized to integrate continuously-weighted evidence maps for determining the best suitable collar locations for borehole testing of identified local-scale targets. Since chromite prospective tracts derived from the geometric average prospectivity model are smaller than those derived from the fuzzy model, the probability of boreholes intercepting chromite mineralization is higher in the geometric average than fuzzy model.