Retrieval-induced forgetting (hereafter, RIF) is defined as the repeated retrieval practice of some materials induces the forgetting of the other materials. Currently, two main theories have been put forward to explain the phenomenon of RIF. The inhibition theory assumes that this forgetting results from the inhibitory processes that function to resolve interference during retrieval. The interference theory proposes that this forgetting is the consequence of interference functioning during the last test phase. Although a bulk of studies was carried out to distinguish these two influential theories, there were still some controversies to be resolved. The present study aimed to examine how the item strength and test order impacted on the RIF. On the basis of this, the debate related with the two theories as mentioned above was discussed in depth. In the present study, sixty-four undergraduates were tested separately. All the forty-eight category-exemplar word pairs were presented once at a time for all the participants. Then, they were instructed to recall half of the exemplars from the half categories which had been studied in the previous phase according to the retrieval cues. By convention, three types of items were produced by the retrieval practice: the practiced exemplars were referred as Rp+ item; the unpracticed exemplars from practiced categories were defined as Rp-item; and the exemplars from unpracticed categories were referred as Nrp item. After an appropriate 2-minute distracter task, participants were shown a category-plus-initial-letter cue and were requested to recall all the exemplars which had been studied before. The results showed that, the retrieval accuracy for the Rp+ items was significantly higher than that for the Nrp items, and this effect was not modulated by the item strength and the test order. Moreover, the RIF was merely observed on the strong items, but not for the weak items. Even when the output effect occurred from the test order were effectively controlled, the strong items still triggered the RIF apparently. Overall, the present results indicated that the effect of RIF was affected by the item strength and the test order. Based on this，we argued that, the inhibitory theory of RIF was supported to some extent.