Structural　health　monitoring　(SHM)　has　been　widely　used　in　all　kinds　of　bridges.　It　is　significant　to　accurately　assess　the　serviceability　and　reliability　of　bridge　subjected　to　severe　conditions　by　SHM　technique.　Bridge　deflection　as　an　essential　evaluation　index　can　reflect　structural　condition　perfectly.　In　this　study,　an　approach　for　deflection　calculation　and　reliability　assessment　of　simply　supported　bridge　is　presented.　Firstly,　a　bridge　deflection　calculation　method　is　proposed　based　on　modal　flexibility　and　Kriging　method　improved　by　artificial　bee　colony　algorithm.　Secondly,　a　dynamic　Bayesian　network　is　employed　to　evaluate　the　deflection　reliability　combined　with　monitoring　results　which　include　modal　frequency,　mode　shape,　environmental　temperature,　and　humidity.　A　linear　regression　model　is　established　to　analyze　the　relationship　between　modal　parameters　and　environmental　factors.　Thirdly,　a　simply　supported　bridge　is　constructed　and　monitored　to　verify　the　effectiveness　of　the　proposed　method.　The　results　reveal　that　the　proposed　method　can　precisely　calculate　the　bridge　deflection.　Finally,　the　time-dependent　reliabilities　of　two　cases　are　computed　and　the　effects　of　monitoring　factors　on　bridge　deflection　reliability　are　analyzed　by　sensitivity　parameter.　It　indicates　that　the　reliability　is　negatively　correlated　with　temperature　and　more　sensitive　to　mode　shape　than　other　three　factors.