Description of precipitation patterns and changes in the hydrological cycle during periods of past global change is crucial for providing an understanding of terrestrial climate systems and for predicting impacts of future climate change such as shifting water availability. While a number of proxies and climofunctions exist for reconstructing paleoprecipitation using paleosols, all of the available tools for reconstructing paleoprecipitation are either limited to certain precipitation ranges (effective only for low-precipitation regimes; e.g., depth to Bk, chemical index of alteration [CIA-K]), or are relevant only to a limited range of paleosols (single-pedotype relationships; e.g., calcium-magnesium index [CALMAG]). Here, we measure the acquisition of isothermal remanent magnetization in B horizons of modern soils to quantify the ratio of pedogenic magnetic minerals goethite and hematite, and we use the relationship between these soil magnetic properties and measured climatic variables at each soil site to derive a new quantitative proxy for precipitation. By compiling both literature-derived and measured goethite-hematite (G/H) ratios and mean annual precipitation estimates for a global suite of modern soils (n = 70), we describe a strong linear relationship (R2 = 0.96) between the G/H ratios of soil B horizons and mean annual precipitation that can be used to estimate paleoprecipitation values for a wide range of climatic regimes (100–3300 mm yr–1) and soil types (Inceptisols, Alfisols, Ultisols, Oxisols, Mollisols, Aridisols, Spodosols). We tested the new climofunction using paleosols from the early Eocene of Wyoming, which show that estimates based on G/H ratios compare favorably to and expand upon previously published estimates based on paleosol data.