Compact　development　aims　to　lead　to　dense　and　mixed　distributions　of　population　and　employment　in　land　use　planning　which　may　result　in　reduced　vehicle　miles　traveled　(VMT).　However,　determining　an　optimum　or　desirable　configuration　of　population　and　employment　distribution　to　achieve　the　compact　development　goal　remains　challenging　due　to　numerous　competitive　development　alternatives　or　possible　exacerbated　traffic　congestion　as　a　result　of　over-intensification　of　urban　land　use.　To　address　this　challenge,　a　systematic　approach　is　proposed　with　a　bi-level　optimization　model　aiming　to　find　out　an　efficient　population　and　employment　map　for　vehicle　travel　reduction.　The　upper　level　of　the　model　is　formulated　to　minimize　VMT　and　vehicle　hours　traveled　(VHT)　varying　with　possible　changes　in　population　and　employment　densities.　The　lower　level　is　based　on　a　tour-based　travel　demand　model　to　mechanistically　represent　the　response　of　travel　choices　to　those　changes.　The　programming　is　solved　by　a　Genetic　Algorithm.　The　proposed　method　is　demonstrated　through　a　case　study　of　Hamilton　County,　Ohio,　U.S.　The　results　indicate　that　a　more　compact　urban　form　reduces　VMT;　however,　it　may　cause　longer　VHT　depending　upon　the　density.　To　avoid　urban　overconcentration　and　reduce　both　VMT　and　VHT　simultaneously,　a　compact　urban　development　configuration　with　a　population　density　＜　5289　persons/mi(2)　and　the　employment　density　＜　3282　jobs/mi(2)　is　recommended　for　the　area.　As　a　negative　relationship　demonstrated　between　transit　headway　reduction　and　vehicle　travel　demand,　strategies　for　improving　transit　provision　are　helpful　to　reducing　vehicle　travels　in　compact　development　scenarios.