We succeed to model biological knowledge on genomic structures to scaffold chloroplast genomes. Our results suggest that modelling genomic regions is sufficient for scaffolding repeats and is suitable for finding several solutions corresponding to several genome forms.