Forests are fragmented, irregular and discontinuous. Therefore, complexity metrics can be used to analyse spatio-temporal dynamics. In this study, the degree of homogeneity or heterogeneity of forests was analysed using lacunarity and the degree of disorder using imaging entropy. The entropy results showed that as deforestation expanded, the entropy of the summed deforested areas increased slightly, indicating that deforestation was taking place in very small areas and mostly along flowing waters. The entropy of the regenerated area was lower than that of the deforested area, indicating that regeneration was relatively more compact and orderly (dominated by plantations). Lacunarity results showed that there was a tendency for heterogeneity to decrease due to clustering of deforestation, but also that regeneration was more heterogeneous than deforestation, with regeneration occurring where deforestation had not occurred. These findings have implications for future forest planning and sustainable management.