Extreme　heat　is　now　a　new　normal　for　many　cities　due　to　climate　change　and　urbanisation.　The　heat　challenges　have　generated　significant　impacts　on　people＇s　daily　functions,　including　driving　behaviours.　In-vehicle　temperature　condition　regulate　drivers＇　thermal　comfort,　responses,　and　thereby　intervening　their　cognition　of　on-board　computer　interface　schemes　(OBCIS).　However,　existing　studies　on　the　OBCIS　usability　evaluation　have　not　well　dealt　with　the　impacts　of　ambient　temperature　and　scheme　presentation　order　on　drivers＇　responses.　This　study　addressed　such　research　gaps　by　empirically　analysing　respondents＇　eye　movement,　including　fixation　duration　(FD)　and　fixation　times　(FT)　towards　24　typical　interface　prototypes　under　six　temperature　conditions.　Moreover,　the　genetic　algorithm　(GA)　and　simulated　annealing　algorithm　(SAA)　were　applied　to　optimise　the　total　distance　between　the　prototypes　for　optimal　OBCIS　presentation　order.　The　results　indicated　that　respondents＇　FD　and　FT　generally　increased　with　temperature.　The　checkerboard　type　OBCIS,　which　got　the　highest　usability　score,　presented　the　lowest　FT　at　23　degrees　C,　while　it　exhibited　the　highest　FT　at　26　degrees　C.　Moreover,　after　the　two　optimisation　algorithms　had　been　run　for　50　times　respectively,　paired　sample　t-test　showed　that　GA　performed　better　than　SAA　in　the　optimisation　work,　and　thereby　the　optimal　presentation　order　was　determined.　This　study　provides　a　reference　for　understanding　heat-induced　information　interaction　and　OBCIS　design.　Compared　with　traditional　usability　evaluation　methods　of　human-computer　interface,　this　research　properly　ensures　the　reliability　of　the　evaluation　results　in　terms　of　ambient　temperature　and　order　effect　bias.