Research in corrective feedback (CF) has developed over thirty years; however, little is known about how metalinguistic corrective feedback (MCF) assists the development of procedural knowledge. This quasi-experimental study sought to evaluate and compare the roles of oral MCF in proceduralizing English as a foreign language (EFL) learners’ declarative knowledge of gender referents and morphosyntactical structure—third person singular -s. Sixty college EFL learners participated in the study. They were randomly assigned to the experimental group and the control group. All learners participated in 14 communicative form-focused activities that elicited the use of gender referents and third person singular morpheme -s in regular classes over 14 weeks. The experimental group received MCF on errors in use of the target structures. The control group received no MCF in use of the target structures. Pretests, posttests, and long delayed posttests administered 20 weeks after the instructional treatment assessed the participants’ declarative knowledge and procedural knowledge of both target structures. The metalinguistic knowledge test was employed to measure declarative knowledge. The elicited oral imitation test was adopted to measure procedural knowledge. Repeated measures ANOVA results indicated that the MCF group significantly increased their procedural knowledge of both target structures. Cohen’s effect size d revealed that MCF exerted a bigger effect on proceduralizing EFL learners’ knowledge on third person singular morpheme –s than on gender referents. These results also demonstrated a relationship between the efficacy of oral MCF and the target structures as well as the interactional effect of test formats tapping into declarative knowledge and procedural knowledge. These findings shed light on the learnability of difficult structures through MCF and the necessity to employ metalanguage in EFL classrooms.