Abstract Background and aims Within-population genetic and phenotypic variation play a key role in the development of adaptive responses to environmental change. Between-population variation is also an essential element in assessing the evolutionary potential of species in response to changes in environmental conditions. In this context, common garden experiments are a useful tool to separate the genetic and environmental components of phenotypic variation. We aimed to assess within- and between-population phenotypic variation of Lupinus angustifolius L. in terms of its evolutionary potential to adapt to ongoing climate change. Methods We evaluated populations’ phenotypic variation of foliar, phenological and reproductive traits with a common garden experiment. Patterns of functional trait variation were assessed with (1) mixed model analyses and coefficients of variation (CVs) with confidence intervals, (2) principal component analyses (PCAs) and (3) correlations between pairs of traits. Analyses were performed at the population level (four populations) and at the latitude level (grouping pairs of populations located in two latitudinal ranges). Key Results Phenotypic variation had a significant genetic component associated with a latitudinal pattern. (1) Mixed models found lower specific leaf area, advanced flowering phenology and lower seed production of heavier seeds in southern populations, whereas CV analyses showed lower within-latitude variation especially in phenological and reproductive traits in southern populations. (2) PCAs showed a clearer differentiation of phenotypic variation between latitudes than between populations. (3) Correlation analyses showed a greater number of significant correlations between traits in southern populations (25 vs. 13). Conclusions Between-population phenotypic variation was determined by contrasting temperature and drought at different latitude and elevation. Southern populations had differential trait values compatible with adaptations to high temperatures and drought. Moreover, they had lower within-population variation and a greater number of trait correlations probably as a result of these limiting conditions, making them more vulnerable to climate change.