We used an isotope dilution method to examine the time-dependent changes in the partitioning and lability of selenium (Se) in a Sri Lankan rice soil after adding fertilizer with selenite (Se(IV)) and selenate (Se(VI)) (1 mg kg−1) and incubation under anaerobic (submerged) (30 days) and subsequent aerobic (drained) conditions (7 days) in controlled reaction vessels mimicking rice paddy water management practices. The Kd (the ratio of sorbed ion to that in solution) values for Se(IV) were significantly (P≤ 0.001) larger than those for Se(VI) in all treatments and at all sampling times. The Kd values for Se(IV) and Se(VI) decreased significantly (P≤ 0.001) with time during the anaerobic and subsequent aerobic phases. Applied Se(IV) fertilizer was rapidly removed into non-labile pools during the anaerobic phase (day 0 = 60% labile and day 14 = no labile Se), with no significant increase in the labile pool following short-term aeration. The results suggest that the rapid decrease in Se(IV) lability may be caused by the strong non-reversible (at least for 7 days) sorption of Se (IV). In contrast, applied Se(VI) fertilizer was 90% labile at 0 day and decreased during the anaerobic phase to 30% after 30 days. There was no significant change in the lability of Se(VI) following the short-term aerobic phase following anaerobic conditions. These results indicate that Se(IV) would not be an effective pre-planting fertilizer for rice production. Selenate is likely to be more effective, but losses to non-labile forms during the submerged phase of rice production may mean that efficiency of pre-planting Se(VI) fertilizer is also compromised.