Abstract: In order to solve the“ping-pong effect”of network switching caused by vehicles moving at relatively high speed，according to the idea of Hidden Naive Bayesian Classification，the relationship between attributes is established by breaking through the assumption that attributes are completely independent in the original Bayesian decision-making． And self-adaptive correction probability is introduced to reduce the number of switching and avoid calculation complexity． The simulation results show that，compared with the original algorithm and other
algorithms，the improved algorithm can effectively reduce the number of handoffs，and has lower running time，
which improves the stability and efficiency of vertical handoff in the environment of vehicular networking．

Key words: the 4 generation mobile communication technology ( 4G) , vehicular ad-hoc network ( VANET) , Wi-Fi, hidden naive bayesian ( HNB) classification, vertical handoff