Abstract. Drought prediction is crucial, especially where the rainfall regime is irregular, such as in Mediterranean countries. A new combined drought indicator (CDI) integrating rainfall, soil moisture and vegetation dynamics is proposed. Standardized precipitation index (SPI) is used for evaluating rainfall trends. A bucket-type soil moisture model is employed for keeping track of soil moisture and calculating anomalies, and, finally, satellite-based normalized difference vegetation index (NDVI) data are used for monitoring vegetation response. The proposed CDI has four levels, at increasing degrees of severity: watch, warning, alert type I and alert type II. This CDI was thus applied over the period 2003–2013 to five study sites, representative of the main grain-growing areas of SW Spain. The performance of the CDI levels was assessed by comparison with observed crop damage data. Observations show a good match between crop damage and the CDI. Important crop drought events in 2004–2005 and 2011–2012, distinguished by crop damage in between 70 % and 95 % of the total insured area, were correctly predicted by the proposed CDI in all five areas.