Previous observational study has demonstrated that spring Arctic Oscillation (AO) has a significant asymmetric impact on the El Ni?o-Southern Oscillation (ENSO) during the following winter. In particular, positive spring AO year can exert a notable impact on the following winter El Ni?o event. However, impact of the negative spring AO on the following winter La Ni?a is weak. In this study, we examine ability of the 30 coupled models participated in the Coupled Model Intercomparison Project Phase 5 (CMIP5) in reproducing the observed asymmetric impact of the spring AO on the ENSO during the following winter. Results showed that only two (i.e. CNRM-CM5 and GISS-E2-H-CC) out of the 30 models can well capture the observed significant spring AO-ENSO relation. These two models were further employed to examine the asymmetric impact of the spring AO on the following winter ENSO. It is showed that the CNRM-CM5 can reasonably well reproduce the observed asymmetric relationship between the spring AO and the winter ENSO. In particular, positive (negative) spring AO could (not) lead to an El Ni?o-like (La Ni?a-like) sea surface temperature (SST) warming in the tropical central-eastern Pacific in the CNRM-CM5. By contrast, the significant impact of the spring AO on the subsequent winter ENSO is symmetry in the GISS-E2-H-CC. Namely, positive (negative) spring AO could result in significant positive (negative) SST anomalies in the tropical central-eastern Pacific during following winter in the GISS-E2-H-CC. The possible factors responsible for the asymmetric/symmetric relation of the spring AO with following winter ENSO in the CNRM-CM5/GISS-E2-H-CC were further examined. For the CNRM-CM5, a significant anomalous cyclone and positive precipitation anomalies can be seen over the subtropical western-central North Pacific during positive spring AO year. The positive precipitation anomalies play an important role in maintaining the westerly wind anomalies over the tropical western Pacific via Gill-type atmospheric response. The westerly wind anomalies over the tropical western Pacific further impact following winter El Ni?o event via triggering eastward propagating and downwelling Kelvin wave. However, during the negative spring AO year, the associated anomalous anticyclone and negative precipitation anomalies over the subtropical North Pacific are weak. Hence, significant easterly wind anomalies cannot be induced over the tropical western Pacific, leading to weak connection of the negative spring AO with the following La Ni?a event. Hence, the CNRM-CM5 can well reproduce the observed asymmetric impact of the spring AO on the following winter ENSO. In comparison, for the GISS-E2-H-CC, significant positive (negative) precipitation anomalies can be seen in the subtropical western North Pacific during the positive (negative) spring AO year, which could induce clear westerly (easterly) wind anomalies in the tropical western Pacific via Gill-type atmospheric response. The resultant westerly (easterly) wind anomalies over the tropical western Pacific further contribute to formation of the El Ni?o (La Ni?a)-like SST anomalies in the tropical central-eastern Pacific during following winter. Hence, the spring AO has a symmetric impact on the following winter ENSO in the GISS-E2-H-CC. Further analysis suggests that the model"s ability in capturing the observed asymmetric impact of the spring AO on the following winter ENSO may be also partly related to the model"s ability in capturing the observed spatial structure of the spring AO.