The　objective　of　this　study　was　to　design　a　universal　method　to　produce　various　degrees　of　mental　workload　and　explore　its　effect　on　driver　reaction　time　through　a　driving　simulator.　Twenty　participants　experienced　a　driving　situation　under　four　degrees　of　mental　workload:　condition　1　(none),　condition　2　(low),　condition　3　(middle),　and　condition　4　(high).　A　program　was　exploited　and　used　to　record　reaction　time　under　each　condition.　A　method　of　mental　arithmetic,　involving　continuous　subtraction　calculations　with　the　same　minuend　and　four　different　subtrahends,　was　introduced　to　produce　differentiable　mental　workload.　Meanwhile,　self-report　was　utilized　to　measure　the　mental　workload.　The　result　of　questionnaires　reflects　that　the　method　of　producing　differentiable　mental　workload　is　feasible　and　universal.　The　data　of　reaction　time　shows　that　the　various　degrees　of　mental　workload　have　a　significant　effect　on　driver　reaction　time,　and　high　mental　workload　leads　to　drivers＇　longer　reaction　time.　The　conclusion　is　that　mental　workload　consumes　the　inherently　limited　capacity　of　driver＇s　attention,　which　leads　to　increase　in　reaction　time.　The　findings　have　a　realistic　significance　on　traffic　safety　research.　©　2013　American　Society　of　Civil　Engineers.