High wetness state levels can be consider ed as a primary indicato r for potential river flooding. Therefore it is advisable to visualise real-time soil moisture information in flood forecasting or warning systems. Monitoring of soil moisture, however, is not an easy task due to its variable nature in time, space and depth. This paper presents and comp ares methods to assess the severity of the soil moisture state of hydrological catchments considered in a typical operational flood forecasting system. The severity of the relative soil moisture state is obtained and mapped by comparing the actual simulation result with the historical simulation results of a lumped concep tual hydrological model, directly by making use of the soil moisture component of the model or indir ectly considering the baseflow component. Another approach uses rainfall, evapotranspiration and river flow observations. By apply ing a baseflow filter to the river flow observations and an advanced method for empirical catchm ent water balance computation, two indirect soil moisture indicators were defined, namely the filtered baseflow and the water balance based relative soil moisture content. It is shown that each of the methods allows to obtain useful esti- mates of the soil moisture state of a catchment in real time. The severity level of the soil moisture state is computed after comparison wit h long term statistics derived from a long term simulation. The seve rity level moreover is used to calculate the probability of exceedance of a predefined riverflow threshold, e.g. flood threshold, at the outlet or a given location in the catchment. This is done by means of a logit relation of the river flow probability of exceedance with the soil moisture indicator. The different soil moisture indicators are compared in their predicting capabilities by calculating and comparing the Brier score. Interestingly, the application of the logit relation or the use of a simple water balan ce computation for the catchment, based on real-time rainfall, evapotranspiratio n and river flow observations, leads to more reliable probability of exceedance estimates than the common direct use of total runoff results from a state-of-the art rainfall–runoff model. Mapping the probability of exceedance for the different hydrolog- ical catchments together with the width of the confidence interval on this probability is proposed as a useful tool to increase the preparedness for potential floods.