The　paper　proposed　a　grey　relation-projection　pursuit　model　for　assessment　of　groundwater　quality.　Grey　correlation　projection　method(GRPM)　with　preferential　membership　and　grey　correlation　has　been　improved,　and　the　improved　grey　correlation　projection　method　was　combined　with　projection　pursuit(PP)　to　present　a　grey　relation-projection　pursuit　model　which　could　be　applied　to　evaluate　the　groundwater　quality　for　incompatibility　and　fuzziness　in　assessment　of　water　quality.　The　comprehensive　assessment　of　groundwater　quality　was　taken　as　a　grey　multiple　objective　decision　making　problem　in　the　model,　then　with　projection　pursuit　method,　the　projection　objective　function　which　could　reflect　groundwater　quality　was　constructed,　and　Firefly　Algorithm(FA)　was　adopted　to　optimize　the　projection　objective　function　for　obtaining　the　best　direction　of　projection　of　grey　correlation　projection　value　and　evaluating　the　grey　relation　projection　value　of　standard　samples　and　evaluation　samples,　thus　the　groundwater　quality　could　be　assessed　comprehensively.　10　random　samples　of　dry　season　of　Beijing(2013)　were　analyzed　with　this　model.　And　a　comparison　of　evaluation　results　from　the　model　and　ones　from　comprehensive　evaluation　method　was　made　and　the　model　has　been　proved　of　being　feasible　and　effective.　The　application　example　showed　that　grey　relation-projection　pursuit　model　could　be　adopted　in　evaluating　the　groundwater　quality　for　its　objectiveness　of　reflecting　the　overall　situation　of　groundwater　quality.　The　model　could　be　applied　in　analysis　and　comparison　of　water　quality　samples　of　same　level　and　provided　reference　in　analysis　of　deteriorating　direction　of　water　quality.　The　evaluation　results　will　be　more　close　to　the　reality.　©,　2015,　The　Editorial　Board　of　Journal　of　Basic　Science　and　Engineering.　All　right　reserved.