In　automated　driving,　takeover　behavior　is　crucial　to　safe　transition　of　control.　Based　on　this　fact,　the　goal　of　the　present　study　is　to　construct　the　driver's　takeover　behavior　mode　and　verify　its　effectiveness.　In　the　takeover　process,　we　define　the　takeover　behavior　mode,　which　is　the　inherent　mode　of　completing　the　takeover　task,　as　the　behavioral　theoretical　framework　that　drivers　should　follow　when　taking　over.　We　conducted　a　driving　simulation　experiment　to　analyze　driver　takeover　behavior　in　the　fog　zone,　where　non-driving-related　tasks　were　external　variables.　First,　we　constructed　the　takeover　behavior　mode　(situation　awareness　-　decision　and　reaction　-　takeover　performance).　Second,　we　analyzed　the　characteristics　of　each　stage　and　the　correlation　between　each　stage　of　the　takeover　behavior　mode　accordingly.　The　result　shows　that　drivers'　performance　in　each　stage　differed　significantly　under　different　non-driving-related　tasks.　For　example,　their　saccade　and　gaze　duration　were　longer　on　the　entertainment　task　than　on　the　work　task,　and　the　driver　prioritized　the　longitudinal　stability　after　taking　control　of　the　automated　vehicle.　Results　show　that　there　is　a　correlation　between　different　stages　of　takeover　behavior　mode.　In　this　study,　we　propose　the　takeover　behavior　mode,　aiming　to　help　beginners　and　researchers　in　this　field　understand　the　driver　takeover　process　more　intuitively　and　provide　behavioral　theoretical　support　for　subsequent　studies.　In　addition,　the　takeover　behavior　mode　can　provide　support　for　training　drivers　in　real-world　driving　situations.　IEEE