Learning with a stimulus from one sensory modality can facilitate subsequent learning with a new stimulus from a different sensory modality. To date, the characteristics and mechanism of this phenomenon named transfer effect still remain ambiguous. Our previous work showed that electrical stimulation of medial prefrontal cortex (mPFC) as a conditioned stimulus (CS) could successfully establish classical eyeblink conditioning (EBC). The present study aimed to (1) observe whether transfer of EBC learning would occur when CSs shift between central (mPFC electrical stimulation as a CS, mPFC-CS) and peripheral (tone as a CS, tone CS); (2) compare the difference in transfer effect between the two paradigms, delay EBC (DEBC) and trace EBC (TEBC). A total of 8 groups of guinea pigs were tested in the study, including 4 experimental groups and 4 control groups. Firstly, the experimental groups accepted central (or peripheral) CS paired with corneal airpuff unconditioned stimulus (US); then, CS shifted to the peripheral (or central) and paired with US. The control groups accepted corresponding central (or peripheral) CS and pseudo-paired with US, and then shifted CS from central (or peripheral) to peripheral (or central) and paired with US. The results showed that the acquisition rates of EBC were higher in experimental groups than in control groups after CS switching from central to peripheral or vice versa, and the CR acquisition rate was remarkably higher in DEBC than in TEBC in both transfer ways. The results indicate that EBC transfer can occur between learning established with mPFC-CS and tone CS. Memory of CS-US association for delay paradigm was less disturbed by the sudden switch of CS than for trace paradigm. This study provides new insight into neural mechanisms underlying conditioned reflex as well as the role of mPFC.