Extinction is an integral part of normal healthy fear responses, while it is compromised in several fear-related mental conditions in humans, such as post-traumatic stress disorder (PTSD). Although much research has recently been focused on fear extinction, its molecular and cellular underpinnings are still unclear. The development of animal models for extinction will greatly enhance our approaches to studying its neural circuits and the mechanisms involved. Here, we describe two gene-knockout mouse lines, one with impaired and another with enhanced extinction of learned fear. These mutant mice are based on fear memory-related genes, stathmin and gastrin-releasing peptide receptor (GRPR). Remarkably, both mutant lines showed changes in fear extinction to the cue but not to the context. We performed indirect imaging of neuronal activity on the second day of cued extinction, using immediate-early gene c-Fos. GRPR knockout mice extinguished slower (impaired extinction) than wildtype mice, which was accompanied by an increase in c-Fos activity in the basolateral amygdala and a decrease in the prefrontal cortex. By contrast, stathmin knockout mice extinguished faster (enhanced extinction) and showed a decrease in c-Fos activity in the basolateral amygdala and an increase in the prefrontal cortex. At the same time, c-Fos activity in the dentate gyrus was increased in both mutant lines. These experiments provide genetic evidence that the balance between neuronal activities of the amygdala and prefrontal cortex defines an impairment or facilitation of extinction to the cue while the hippocampus is involved in the context-specificity of extinction.