In order to provide a relatively simple means of predicting live herbaceous plant moisture content from a readily available meteorological index with an accuracy adequate for fire hazard assessment, the moisture content fluctuations of certain species were correlated with the values of a seasonal drought index based on soil moisture deficiency. The simple linear regression models provided the best fit of the relationship between plant moisture content and the Keetch-Byram drought index (KBDI) values. KBDI predicted with accuracy, for two growing seasons, the moisture content of three annual herbaceous plants ( Piptatherum miliaceum, Parietaria diffusa, Avena sterillis) with shallow rooting systems, typical of the understory vegetation of Pinus brutia forests in the Mediterranean region of Crete, Greece. The greatest aberrations between measured and predicted values of plant moisture content were observed early (May) and late (September) in the growing season, when plant phenology (flushing and withering stages respectively) appears to become the dominant factor in determining plant moisture content regardless of the soil moisture conditions. The KBDI was poorly correlated with the live-needle moisture content of deeply rooted P. brutia trees and modestly with the soil water content of the upper layers. This indicates that the index adequately reflects the moisture condition of the surface soil layers but not the water content deeper in the soil.