The age of some lava flows on the east-north-east coast of Fogo Island (Cape Verde) remains debated or unknown due to the inapplicability of conventional dating methods. Dating information is important for a retrospective approach to volcanic hazard evaluation and prevention. Historical lavas represent an opportunity to study plant succession within a well-constrained chronological framework, and lava flow age has often been used as a variable to explain plant densities and species composition. Here, vascular plant densities were also assigned as variables to explain lava ages. We surveyed 44 plots on eleven semi-arid lavas among which nine represented a chronosequence of 60- c. 350-yr-old lavas, while two other lavas remained undated. We used principal component analysis (PCA) to test the relationships between species densities and lava ages. A stepwise regression model was then used to identify which species were more closely related to the age of lavas, and the best fit of the model was employed to predict the age of the undated lavas. Spearman' test was used to observe the correlations among lavas and the result was compared with regression model age-predictions for the undated lavas. Among 27 vascular plants, nine species were significantly linked with lava ages. The stepwise model showed that lava ages were significantly explained by the densities of Sarcostemma daltonii and Lantana camara species. Stepwise model age-predictions were congruent with Spearman' results, supporting the proposition that vascular plant species densities can provide useful information about the relative age of undated lava flows in arid environments.