Geostatistical methods to predict soil properties using high-quality geospatial datasets have been explored for some time, but it has not been until the last decade, with advances in geographic information systems and remote sensing, that these methods have been improved to incorporate secondary or ancillary environmental variables into the mapping of soil properties. These have been defined as “hybrid interpolation techniques”. Our objective was to compare univariate and hybrid geostatistical methods to predict soil total phosphorus (TP) in Water Conservation Area-2A, Everglades, Florida. Soil data were collected at 111 sites at 0-10 cm depth in 2003. We used ordinary kriging, co-kriging and regression kriging to predict TP. To conduct the hybrid modeling auxiliary environmental variables were derived from Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) NASA satellite imagery. The hybrid geostatistical methods performed better to predict TP than ordinary kriging. Our findings indicate that auxiliary, dense datasets derived from remote sensing images have value to improve the predictions of soil TP in wetland systems.