Experiments　were　conducted　with　a　driving　simulator　to　examine　how　different　blood　alcohol　concentrations　(BACs)　and　fatigue　levels　influenced　driver　safety　performance　in　various　roadway　geometries.　Each　of　the　22　participants　completed　five　drives.　The　control　condition,　serving　as　the　benchmark　in　the　five　drives,　refers　to　driving　under　sober　conditions.　The　four　testing　conditions　included　drives　with　three　BAC　levels　(0.02%,　0.05%,　and　0.08%)　and　drives　in　a　fatigued　state.　The　experimental　test　scenarios　covered　both　straight　and　curved　roads.　The　tangent　straight　roadway　segments　were　used　to　connect　the　curved　ones　of　three　different　radii　(200,　500,　and　800　m)　in　both　travel　directions.　Statistical　analysis　of　result　variance　was　used　to　quantify　the　effectiveness　of　measures　of　vehicle　longitudinal　speeds　and　lane　positions.　The　averages　and　standard　deviations　of　speeds　and　lane　positions,　indicated　as　SP_AVG,　SP_SD,LP_AVG,　and　LP_SD,　were　all　significantly　affected　by　drivers＇　states.　Experimental　test　results　indicated　that　different　levels　of　BAC　and　fatigue　imposed　different　influences　on　driver　safety　performance.　Road　geometric　characteristics,　such　as　radii　and　curvature　directions,　had　significant　impacts　on　drunk　and　fatigued　drivers＇　capability　of　manipulating　vehicles　for　safe　operations.　The　research　findings　provide　reference　information　to　identify　hazardous　driving　status,　such　as　drunk　and　fatigued　driving　under　different　geometric　conditions.