Fear extinction is not permanent but more vulnerable than the original fear memory, as relapse phenomena have traditionally shown. Partial extinction has been proposed as a strategy that may serve to mitigate relapses. Partial extinction differs from the standard procedure as it includes the occasional presentation of CS-US trials at the beginning of the extinction training. The present experiment, using an aversive differential conditioning procedure, evaluates whether partial extinction can reduce reinstatement, a specific form of relapse. The results showed that partial extinction did not mitigate reinstatement but proved effective to diminish the magnitude of the US expectation after a first reacquisition trial in a final test phase. The results reported are more consistent with theories that conceives extinction as the acquisition of new inhibitory learning rather than the erasure of the original conditioning.