VOCs,　as　the　common　precursor　of　PM2.5　and　O3　pollution,　has　not　been　paid　enough　attention　in　the　previous　phase.　How　to　implement　scientific　and　effective　emission　reduction　on　VOC　sources　is　the　focus　of　the　next　step　in　improving　the　atmospheric　environmental　quality　in　China.　In　this　study,　based　on　observations　of　VOC　species,　PM1　components　and　O3,　the　distributed　lag　nonlinear　model　(DLNM)　was　used　to　investigate　the　nonlinear　and　lagged　effects　of　key　VOC　categories　on　secondary　organic　aerosol　(SOA)　and　O3.　The　control　priorities　of　sources　were　determined　by　com-bining　the　VOC　source　profiles,　which　were　afterwards　verified　using　the　source　reactivity　method　and　Weather　Re-search　and　Forecasting　Model-Community　Multi-scale　Air　Quality　Model　(WRF-CMAQ).　Finally,　the　optimized　control　strategy　of　VOC　source　was　proposed.　The　results　showed　that　SOA　was　more　sensitive　to　benzene　and　toluene,　and　single-chain　aromatics,　while　O3　was　more　sensitive　to　dialkenes,　C2-C4　alkenes,　and　trimethylbenzenes.　The　op-timized　control　strategy　based　on　the　total　response　increments　(TRI)　of　VOC　sources　suggests　that　passenger　cars,　in-dustrial　protective　coatings,　trucks,　coking,　and　steel　making　should　be　considered　as　the　key　sources　for　continuous　emission　reduction　throughout　the　year　in　the　Beijing-Tianjin-Hebei　region　(BTH).　Non-road,　oil　refining,　glass　manufacturing　and　catering　sources　should　be　strengthened　in　summer,　while　biomass　burning,　pharmaceutical　manufacturing,　oil　storage　and　transportation,　and　synthetic　resin　need　more　emphasis　in　other　seasons.　The　multi-model　validated　result　can　provide　scientific　guidance　for　more　accurate　and　efficient　VOCs　reduction.