Wetlands play an integral role as natural climate solutions due to their potential to sequester carbon dioxide (CO₂) from the atmosphere. However, myriad factors influence the spatial heterogeneity of organic carbon (OC) sequestration rates, necessitating techniques to reduce uncertainty in these rates if they are to guide policy decisions to meet national climate action targets. In this study, we combined expert knowledge with statistical learning techniques to derive place-based estimates of OC sequestration rates for wetlands on agricultural landscapes. Expert knowledge revealed complex relationships between process controls and OC sequestration rates, including carbon quantity, carbon quality, bulk density, cation exchange capacity, aggregate reactivity, redox potential, and temperature for wetlands. GIS and remotely sensed data were used as proxies for these process controls, which, along with field-based OC sequestration rates, were input into a statistical learning-based random forest (RF) model. This RF model predicted OC sequestration rates within reasonable error bounds, achieving adjusted coefficient of determination (R²) of 0.93, a mean absolute error of 0.05 Mg ha-1 yr-1, and a root mean square error of 0.08 Mg ha-1 yr-1. The variable importance analysis indicated that human impact index, various soil properties, and inundation probability were the most influential variables. Our findings suggest that statistical learning-based models grounded in an understanding of process controls and their interactions can reliably estimate OC sequestration rates, providing essential data to inform policy development and implementation for managing wetlands as natural climate solutions.