Experimental loss of occlusal support caused by the extraction or grinding of molar teeth has been reported to foment the impairment of learning and memory in laboratory animals. The purpose of this study was to examine the effect of occlusal reconstruction after long-term loss of molars on spatial memory by using 8-arm radial maze and by assessing histopathological changes of neuron density in the hippocampus. Experimental dentures were inserted into the oral cavities of molarless rats to recover the occlusal support. Age-matched groups of control, molarless and denture-wearing rats were trained to perform the maze tasks. The difference of the error incidence in the maze task was evaluated between three groups. The difference of neuron density between three groups was also evaluated at the end of the maze task. Serum corticosterone levels were also measured to estimate the chronic stress, which could be caused by extraction, insertion of the experimental denture or any experimental procedure. The error incidence in the denture-wearing group was significantly higher than that of the control group, but significantly lower than that of the molarless group. Significant differences of neuron density were observed between three groups in each of the hippocampal CA1, CA3 and DG subfields. No significant difference of the serum corticosterone levels between three groups could be observed. From the results of this study, it was suggested that the recovery of occlusal support would bring amelioration of cognitive impairment concomitant with long period loss of molars in rats.