In　this　study,　a　driving　simulator　experiment　was　conducted　to　investigate　the　influence　of　tunnel　retro-reflective　arch　(TA)　on　driving　behavior,　through　the　employment　and　analysis　of　TAs　with　disparate　spacing　alternatives　(no　TAs,　TAs　with　a　spacing　of　200　m,　and　TAs　with　a　spacing　of　300　m)　and　tunnel　lengths　(800　m,　1800　m,　3500　m,　and　5800　m).　The　performances　of　a　total　of　32　drivers　were　collected　and　assessed.　The　tunnels　and　TAs　were　accurately　replicated　in　the　driving　simulator.　On　the　basis　of　the　experimental　data,　four　behavioral　variables,　namely,　average　speed,　the　standard　deviation　of　speed,　the　standard　deviation　of　acceleration,　and　lateral　position　were　obtained.　A　series　of　linear　mixed　models　with　random　effects　were　estimated　to　reveal　the　contributing　factors　of　driving　behavior　subjected　to　TAs　variation.　The　results　demonstrate　that　drivers　showed　significantly　higher　consciousness　of　speed　control　and　lower　mental　stress,　coupled　with　less　risky　driving　behavior,　under　the　condition　of　TAs　in　long　and　extra-long　tunnels.　Furthermore,　drivers　are　more　comfortable　subjected　to　TAs　with　a　spacing　of　300　m.