In this study, we aimed to clarify the effect of predatory mites on controlling the rubber tree pest mite Eotetranychus sexmaculatus (Riley) and explored basic application of the biocontrol technology “using predatory mites to control pest mites” to rubber trees. We selected two commercial species with high predation capacity from five species of predatory mites for further field-release tests. The results showed that among the five predatory mite species, Phytoseiulus persimilis (Athias-Henriot) and Neoseiulus cucumeris (Oudemans) showed the highest and lowest daily average predation amounts on E. sexmaculatus, respectively, and that the commercial predatory mites Amblyseius swirskii (Athias-Henriot) and Neoseiulus barkeri (Hughes) were demonstrated relatively high predation capacities on E. sexmaculatus and represent the ideal species for field release. Additionally, we found that field release by a hanging release sachet was suitable for a low field-population density of pest mites, and that predatory mites could be released at a dose of one sachet per tree and/or every other row or every other tree according to the actual situation of the pest mites. If the field population of E. sexmaculatus is at the peak egg-laying period, the release of A. swirskii should be top priority, which will lead to high control efficiency within 2 to 3 weeks. In case of low field-population density of E. sexmaculatus or a rainy season, top priority should be given to the release of N. barkeri, which might lead to high control efficiency within 3 to 4 weeks. The results showed that the release of A. swirskii had a more fast-acting effect with a higher control efficiency than N. barkeri at the early stage of release, whereas N. barkeri exhibited a better release effect with greater persistence than A. swirskii at the late stage of release.