We present the first study using decadal re-forecasts to simulate global glacier climatic mass balance, bridging the gap between seasonal and long-term simulation of glacier contribution to catchment hydrology and sea level rise. Using the Open Global Glacier Model, driven by Coupled Model Intercomparison Project 6 ensembles of initialised decadal climate re-forecasts of temperature and precipitation, we demonstrate the skill of glacier mass balance re-forecasts on the decadal timescale, for respectively 279 reference glaciers and all land-terminating glaciers globally. For comparison, the glacier model is also forced with a simple persistence forecast and general circulation model historical time series and projections, representing the current state of the art. The results from forcing with decadal re-forecasts provide improvement over the other two methods. Simulating single years, especially at short lead times, decadal re-forecasts show the highest Pearson correlations and lowest mean absolute errors, compared to observed mass balance. Simulating cumulative mass balance over full decades for the 279 reference glaciers, forcing with decadal re-forecasts yields a decrease in mean absolute error of 18 % and 16 % compared to forcing with persistence forecasts and historical global circulation model simulations, respectively. Globally, comparing average mass balance over the time period 2000–2020, forcing with decadal re-forecasts results in the highest number of regions with ’good fit’ to observations (difference from observed regional mass balance =< 0.1 m w.e.), compared to the persistence and historical climate model forcing. These findings indicate that the use of decadal predictions for glacier modelling is operationally feasible and holds significant potential for future hydrological applications.