Testing one's memory of previously studied information reduces the rate of forgetting, compared to restudy. However, little is known about how this direct testing effect applies to action phrases (e.g., "wash the car") - a learning material relevant to everyday memory. As action phrases consist of two different components, a verb (e.g., "wash") and a noun (e.g., "car"), testing can either be implemented as noun-cued recall of verbs or verb-cued recall of nouns, which may differently affect later memory performance. In the present study, we investigated the effect of testing for these two recall types, using verbally encoded action phrases as learning materials. Results showed that repeated study-test practice, compared to repeated study-restudy practice, decreased the forgetting rate across 1 week to a similar degree for both noun-cued and verb-cued recall types. However, noun-cued recall of verbs initiated more new subsequent learning during the first restudy, compared to verb-cued recall of nouns. The study provides evidence that testing has benefits on both subsequent restudy and long-term retention of action-relevant materials, but that these benefits are differently expressed with testing via noun-cued versus verb-cued recall.