Electric　vehicles　(EVs)　have　been　developed　rapidly　in　recent　years　in　China,　but　a　further　promotion　of　EVs　is　hindered　by　the　unreasonable　deployment　of　facilities.　Therefore,　a　deep　understanding　of　EV　drivers＇　charging　behavior　is　viewed　to　be　helpful　to　a　more　suitable　deployment　of　new　chargers.　In　this　paper,　charging　influence　factors　are　analyzed,　and　logit-based　models　for　charging　choice　behaviors　are　proposed.　First,　a　web-based　stated　preference　survey　is　designed　to　obtain　the　data　of　EV　drivers＇　charging　choice　preference.　Second,　binary　logit　models　are　developed　using　the　survey　data,　with　vehicle　attributes,　destination　activity,　and　subsequent　travel　information　as　influence　factors.　Especially,　the　satisfaction　of　charging　facilities　reviewed　by　previous　users　and　drivers＇　risk　attitudes　are　also　introduced　to　analyze　charging　choice　preference.　Third,　latent　class　models　are　proposed　to　further　taste　the　heterogeneity　of　EV　drivers＇　charging　decision　strategies.　The　results　reveal　two　classes　of　decision　making　patterns　among　our　sample　population:　(1)　service　concerned,　which　are　the　majority　of　the　respondents　who　value　the　satisfaction　and　queue　of　charging　station,　characterized　with　younger　age,　richer　driving　experience　and　higher　income,　and　(2)　pragmatic　concerned,　those　who　weight　multiple　factors,　such　as　excess　range,　parking　time,　and　charging　fee.　It　suggests　that　the　improvement　of　charging　service　levels,　such　as　providing　reliable　and　high-quality　charging　service,　is　necessary　for　operators　to　attract　more　EV　drivers.　(c)　2020　Elsevier　Ltd.　All　rights　reserved.