Logged-over forests comprise a large proportion of tropical forest landscapes, but questions remain on how to fully describe and understand the regeneration of these forests and which selective logging and silvicultural interventions are best for maintaining biological diversity. We used a phylogenetic approach to capture changes on the evolutionary diversity, structure and composition of 72-one-hectare forest subplots in East Kalimantan that were logged using conventional (CL) and reduced-impact (RIL) techniques, including post-logging silvicultural interventions by reducing stem density (thinning), shortly after (on average 18 months) and 10 years after treatments. Both CL and RIL techniques initially decreased phylogenetic diversity, but after a decade of succession the phylogenetic diversity levels were no longer different from unlogged forests (Controls). Thinning through poison-notching non-commercial trees led to an increase in phylogenetic diversity after a decade of succession. Mean evolutionary relatedness, as measured by the Net Relatedness Index, initially decreased in logged forests, possibly because of harvest of closely related dipterocarp species, but returned to initial levels after a decade. Regardless of silvicultural interventions 10 years after logging had no effect on the mean evolutionary relatedness of the trees in the communities at both time intervals. Both CL and RIL techniques increased the phylogenetic composition dissimilarity with unlogged forest, with increasing logging intensity resulting in increasingly larger effects, although the changes were smaller for RIL. Phylogenetic composition only became dissimilar from unlogged forest after a decade of succession, when the logging intensity was above 4 m2 ha−1 for both CL and RIL, equivalent to 8 trees ha−1 or 68 m3 ha−1. Post-logging silvicultural interventions started to cause changes in phylogenetic composition when the basal area was reduced by thinning by more than 10 m2 ha−1. The rapid recovery of phylogenetic diversity, community structure and, to a lesser extent composition after logging and thinning confirms that tropical forests are resilient to a single logging cycle. With low-intensity logging of less than 4 m2 ha−1, RIL techniques appear to facilitate phylogenetic structure and compositional recovery to levels not different from unlogged forest after 10 years of succession, especially when combined with post-logging silvicultural intervention with a basal area reduction by thinning of less than 10 m2 ha−1. However, it remains to be seen whether forests will respond similarly to a second logging cycle.