This study focused on simulating the long-term soil carbon sequestration after application of anaerobically digested and non-digested cattle manure using the Daisy model. The model was parameterized and calibrated for soil carbon (C) release during a 247 days incubation study including a coarse sand and a loamy sand amended with the two manures. The calibrated set-ups used to simulate long-term cereal cultivation with annual C application of 1023 kg C ha^-1 in non-digested manure or 543 kg C ha^-1 in digested manure (similar manure N application). Long-term simulations were made with annual application of the two manures (70 kg organic manure N ha^-1 plus 90 kg mineral N ha^-1) and compared with a mineral N reference (120 kg N ha^-1 yr^-1). Carbon retention in soil was related to the initial C in non-digested manure, and after 52 years of repeated manure application extra C retention was equivalent to 41% for non-digested and 35% for digested manure in the loamy sand. In the sandy soil corresponding C retention was 37 and 29%. The higher C retention from non-digested compared to digested manure differed from the incubation study and was mainly due to the model response to the optimized parameters for the turnover of the two types of manures. The simulated increase in soil C with cereal crop and also without application of organic manure in the sandy loam is in contrast to previous long-term field studies at soil from the same location in the Askov. This dissimilarity was mainly due to slow degradation of crop residual C in the inert pool during calibration. The results indicate the need for additional calibration of crop residual C and show that calibration based on short-term experiments might not be applicable for estimating the long-term response to C application.