Camelina is an alternative oilseed crop species with limited information about the origin and diversity of available germplasm. Therefore, a set of 130 camelina accessions from a world collection was evaluated for oil content, protein content and 1000-seed weight in field experiments grown in three macro-environments in Austria. Based on phenotypic data, accessions were categorized into four groups with different seed characteristics using k-means cluster analysis or principal component extraction. Subsequently, a representative set of 41 accessions was subjected to random amplified polymorphic DNA (RAPD) analysis. Of 24 primers, 15 were polymorphic producing a total of 30 marker loci. Genetic distance estimates between the 41 accessions were calculated, based both on RAPD polymorphism and on seed quality characteristics, and dendrograms were generated for comparison. Similarities were found between the two different clustering approaches, and grouping was partly in agreement with pedigree information or geographic origin. However, as the two estimates of diversity sampled different segments of the genome, i.e. regions coding for seed characteristics or phenotypically neutral genomic regions highlighted by discrete markers, the correlation between the two distance matrices was low.