The rapid expansion of perennial crops is a major threat to biodiversity in Southeast Asia. The biodiversity losses related to the conversion of forest lands to oil palm or rubber plantations (RP) are well documented by recent studies. However, the impact of the conversion from intensively managed annual crops to perennial crops on soil biodiversity has not yet been addressed. This study aims at assessing the impact on soil biodiversity of a) the short-term effect of land use conversion from cassava crop to RP, and b) the long-term effect of RP ageing. Soil biodiversity (bacterial, fungal and macrofaunal), microbial activities and pedoclimatic characteristics were measured over a chronosequence of 1–25 years old of RP compared to cassava fields, the former crop, in Thailand. The conversion from cassava to young RP (1–3 yr) had a significant effect on microbial biomass and activities and fungal composition, but did not impact the bacterial and macrofaunal diversity. This effect of land use conversion could be explained by the change in land management due to the cultivation of pineapple in the inter-row of the young RP. Canopy closure appeared to be the main driver of soil biota shifts, as most of the biotic parameters, composition, abundance and activities were significantly modified after 7 years of RP. The changes of composition in older rubber plantations originated from the dominance of Trichoderma (fungi), Firmicutes (bacteria), and earthworms. Old rubber plantations (23–25 yr) harboured the highest microbial and macrofaunal biomass; however, they were also characterised by a significant decrease in bacterial richness. The change in pedoclimatic conditions across the rubber chronosequence, i.e. increase in soil moisture, litter and organic carbon content, was a stronger driver of soil biota evolution than land use conversion. The macrofaunal composition was more resistant to land use conversion than the bacterial composition, whereas the microbial biomass was sensitive to land use conversion, but showed resilience after 20 years. However, bacterial, fungal and macrofaunal diversity, macrofaunal and microbial biomass and microbial activities were all sensitive to RP ageing.