The objectives of the present study were to explore the potential of soil invertebrate taxa (families or genera) to evaluate forest management practices. An experiment with four treatments: control, understory removal, Cassia alata (a legume shrub) addition, and both understory removal and C. alata addition, was conducted at Heshan Hilly Land Interdisciplinary Experimental Station, Guangdong, China. Redundancy analysis showed that some bacterivores and fungivores of soil invertebrates (nematodes, mites, collembolans) were correlated positively, but some predators and omnivores were correlated negatively with forest management practices. C. alata addition increased the abundance of the high trophic-level nematodes and mites, which indicated C. alata addition was a good forest management practice in terms of improved soil food web structure. In contrast, removal of forest understory plants appeared to disturb the ecosystem by suppressing high-trophic groups of soil invertebrates, demonstrating that understory removal was not a good forest management practice. Redundancy analysis also showed that soil fauna at the genus or family level can be used as biological indicators for forest management practices. Specifically, some high trophic-level nematode genera such as Eudorylaimus , Chrysonema , Iotonchus and Thornia were suppressed significantly by understory removal and some nematode genera such as Prismatolaimus , Eudorylaimus , Chrysonema , and Thornia and one common mite genus Rhodacarus in high trophic-level were enriched significantly by legume addition.