Abstract Pulsing storms and prolonged rainfall events have been associated to floods, soil erosion and nutrient fluxes in many European river catchments. This motivated us to develop a parsimonious approach to model the climate forcing on sediment yields in a mountainous Austrian–German river catchment. The hydro-climatologic forcing was interpreted by the novel RAMSES (RAinfall Model for SEdiment yield Simulation) approach to estimate the annual sediment yields. We used annual data on suspended-solid yields at the gauge Fussen, monitored from 1924 to 2003, and monthly rainfall data. The dataset was split into the period 1924–1969 for calibration and the period 1970–2003 for validation. The quality of sediment yield data was critically examined, and a few outlying years were identified and removed from further analyses. These outliers revealed that our model underestimates exceptionally high sediment yields in years of severe flood events. For all other years, the RAMSES performed well against the calibration set, with a correlation coefficient (r) equal to 0.83 and a Nash–Sutcliffe Index (NSI) of 0.69. The lower performance in the validation period (r = 0.61, NSI = 0.36) has to be partly attributed to discontinuities in the monitoring strategy. For the calibration dataset, monthly precipitations proved nonetheless to be better predictors for annual sediment yields than annual values. These first results lay the foundation for reconstructing intra- to inter-decadal variability of sediment yields in river catchments where detailed precipitation records are not available as well as for the reconstruction of historical sediment yields.