Abstract Four Sri Lankan isolates of an entomopathogenic fungus, Hirsutella thompsonii , obtained from the coconut mite Aceria guerreronis , were tested as biopesticide treatments in two coconut mite-infested coconut estates. Two applications of a spore suspension, 2 weeks apart, were carried out on the five youngest coconut bunches, including the inflorescence. Two weeks following the second application the level of coconut mite infestation was lower in three of the four H. thompsonii treatments compared to controls, indicating a relatively quick impact on the coconut mite population. Less than 10% of the nuts receiving isolate IMI 391722 had high levels of coconut mites (>100 live mites) at 4 weeks, comparing favourably with all other treatments. As the sampling time progressed from 4 weeks until the completion of sampling at 30 weeks, efficacy was less consistent. However, with the exception of one sample date, nuts treated with IMI 391722 had a lower percentage of nuts with >100 live coconut mites than the control nuts. A significant proportion of nuts with dead coconut mites showing mycosis due to H. thompsonii was recorded for all isolates up to 6 weeks following the second application, with evidence of mycosis up to 18 weeks for isolates IMI 391722 and IMI 390486, and up to 14 and 10 weeks for IMI 391723 and IMI 391942, respectively. Nuts with dead coconut mites showing mycosis were rare on untreated palms throughout the study. Numbers of the predatory mite Neoseiulus baraki were significantly higher on the untreated nuts than the treated nuts. However, dead mites were rarely seen and there were no signs of H. thompsonii infection.