The　traditional　Vehicular　Ad-hoc　Network　(VANET)　routing　algorithms　are　just　designed　based　on　the　scenario　of　a　single　service　traffic　type　and　RAT,　without　designing　personalized　routing　schemes　for　different　types　of　service　traffic.　Software　Defined　Vehicular　Network　(SDVN)　paradigm　can　solve　these　problems　with　its　centralized　control　feature.　We　propose　an　intelligent　SDVN　architecture　for　the　VANET　scenarios　with　different　types　of　service　traffic　and　various　RATs.　Then　we　propose　a　hierarchical　intelligent　routing　algorithm　based　on　Q-learning　considering　the　different　personalized　needs　of　various　traffic　types.　The　simulation　results　show　that　the　algorithm　can　meet　the　personalized　requirements　of　various　traffic,　and　has　better　performance　in　delay-packet　loss　rate-cost.　©　2021　IEEE.