Carbon stored in coastal wetland ecosystems is of global relevance to climate regulation. Broad-scale inventories of this "blue" carbon store are currently lacking and labour intensive. Sampling 23 salt marshes in the United Kingdom, we developed a Saltmarsh Carbon Stock Predictor (SCSP) with the capacity to predict up to 44 % of spatial variation in soil organic carbon (SOC) from simple observations of plant community and soil type. Classification of soils into two types (sandy or not-sandy) explained 32 % of variation in SOC. Plant community type (5 vegetation classes) explained 37 % of variation. Combined information on soil and plant community types explained 44 % of variation in SOC. GIS maps of SOC were produced for all salt marshes in Wales (~ 4000 hectares), using existing soil maps and governmental vegetation data, demonstrating the application of the SCSP for large-scale predictions of blue carbon stores and the use of plant community traits for predicting ecosystem services.