Here we test the predictions of succession in a restoration context to ascertain the validity of succession-based management. Our study took place in seven coastal dune forest sites of various ages regenerating after mining disturbance. We used our 16 yr of data on 4 taxa to test the predictions of succession theories. Patterns in turnover for all taxa showed a decelerating decrease contradicting Clements’s classical theory of succession. Changes in composition followed patterns predicted by the individualistic model of succession. Trends in species diversity measures did not always match predictions with bird species diversity declining in most sites over time. Regional disturbances may explain this phenomenon. Succession-based management is a valid approach to dune forest rehabilitation as long as restoration managers recognise disturbance as an ecological reality.