Since the passage of Public Law 95-87, the Surface Mining Control and Reclamation Act (SMCRA) in 1977, reclamation success of prime farmland after coal mining has been determined by long-term crop yield testing. States such as Illinois and Indiana require that reclamation success be based on crop production of mined land. This process often can continue for many years, especially for lands failing to meet production standards in a specified time period. Needs have been expressed by landowners, mine operators, and regulators for methods to expediate this process. A soil property based model could relieve this burden and ensure the most efficient process for returning the soil resource to the landowner. The objective of our work was to develop a soil-based model to replace the current crop yield-based system and to evaluate mined-land for diagnostic purposes. Geo-referenced corn (Zea mays L.), soybean [Glycine max (L.) Merr.], and wheat (Triticum aestivum L.) yield, cone penetrometer test (CPT), VIS-NIR spectrophotometer, apparent electrical conductivity (ECa), elevation and terrain derivatives, fertility, and other site characteristic data were collected on fields at the Lewis Mine site in southwestern IN, the Cedar Creek Mine site in western IL, and the Wildcat Hills Mine site in southern IL. Soil-based productivity models were developed using regression and multivariate techniques to assign probabilities of meeting crop yield standards at the partial-field level. Our research indicates that soil compaction and water availability primarily influence a field's ability to meet crop yield standards across time. Model validation between fields and among sites has been encouraging, thus we propose modeling soil variability as a diagnostic tool to identify problematic field areas and to complement yield-based requirements.