The extinction of threat expectancies is a fragile phenomenon, as these responses may reappear later. In recent years, several techniques have been proposed to mitigate relapse. One of them is the deconditioning treatment, in which the intensity of the US is gradually reduced across the extinction session so that by its end the US will be still presented but will have no relevant arousing effect. In this study, we tested the effect of this extinction procedure to prevent spontaneous recovery in comparison to a standard extinction treatment. Participants first learned the relationship between a figure and an electric shock using a differential conditioning paradigm and were then randomly assigned to one of the two groups. Finally, after a 20-min retention interval, spontaneous recovery was tested. By and large, we did not find evidence of any relapse prevention produced by the deconditioning treatment when compared to standard extinction. Our results are discussed in relation to recent evidence and different theoretical accounts regarding extinction learning.