Drosophila subobscura presents a rich and complex chromosomal inversion polymorphism. It can thus be considered a model system i) to study the mechanisms originating inversions and how inversions affect the levels and patterns of variation in the inverted regions, and ii) to study adaptation at both the single-gene and chromosomal inversion levels. It is therefore important to infer its demographic history since previous information indicated that its nucleotide variation is not at mutation-drift equilibrium. For that purpose, we sequenced 16 non-coding regions distributed across those parts of the J chromosome not affected by inversions in the studied population and possibly either by other selective events. The pattern of variation detected in these 16 regions is similar to that previously reported within different chromosomal arrangements, suggesting that the latter results would, thus, mainly reflect recent demographic events rather than the partial selective sweep imposed by the origin and frequency increase of inversions. Among the simple demographic models considered in our ABC analysis of variation at the 16 regions, the model best supported by the data implies a population size expansion soon after the penultimate glacial period. This model constitutes a better null model and it is therefore an important resource for subsequent studies aiming among others to uncover selective events across the species genome. Our results also highlight the importance of introducing the possibility of multiple hits in the coalescent simulations with an outgroup.