Harvest residues contain large stores of carbon (C) and nitrogen (N) in forest plantations. Decomposing residues can release labile C and N into soil and thus provide substrates for soil bacterial communities. Previous studies showed that residue retention could increase soil C and N pools and activate bacterial communities in the short term (≤ 10 years). The current study examined the effects of a long-term (19-year) harvest residue retention on soil total and water and hot water extractable C and N pools, as well as bacterial communities via Illumina MiSeq sequencing. The experiment was established in a randomised complete block design with four replications, southeast Queensland of Australia, including no (R0), single (R1, 51 to 74 t ha−1 dry matter) and double quantities (R2, 140 t ha−1 dry matter) of residues retained. Generally, no significant differences existed in total C and N, as well as C and N pools extracted by water and hot water among the three treatments, probably due to negligible amounts of labile C and N released from harvest residues. Soil δ15N significantly decreased from R0 to R1 to R2, probably due to reduced N leaching with residue retention (P < 0.001). Residue retention increased the relative abundances of Actinobacteria (P = 0.016) and Spartobacteria (P < 0.001), whereas decreased Betaproteobacteria (P = 0.050). This favour for the oligotrophic groups probably caused the decrease in the bacterial diversity as revealed by Shannon index (P = 0.025). Hence, our study suggests that residue retention is not an appropriate management practice in the long term.