Inventories of peat volume and carbon storage often include general values for peat depth, but more spatially explicit and accurate estimates are required if carbon management strategies are to be developed at scales appropriate for the management. This article presents a methodology for estimating peat depth for large blanket peat areas using field sampling and GIS modelling to map peat depth on Dartmoor in south‐west England. The study area was divided into carbon unit areas (CUAs) based on soil and vegetation. Approximately 1000 peat depth measurements were taken, each consisting of a mean (n = 5) from depths within a 32 m2 area. Sampling points were stratified according to CUA area and proportional extent of slope and elevation classes. Regression analyses were used to determine the relationships between slope, elevation and peat depth within each CUA. The strongest relationship was for blanket peat (r2 = 0.53), with weaker ones for areas where peat was shallow and depth was less variable. A digital elevation model was used in a GIS to model peat depths for the whole of Dartmoor. Results were tested against a data set of 200 peat depths on a 250 m grid covering 1325 ha. We conclude that peat depth can be modelled using easily available topographic data combined with well‐designed field sampling over larger spatial scales. The approach can result in accurate mapping of peat depth and carbon storage for blanket peatlands in the United Kingdom and perhaps also elsewhere.