The　rational　allocation　of　water　resources　has　become　an　important　prerequisite　for　the　protection　of　national　economic　and　social　sustainable　development　as　the　China　economic　develops　rapidly,　therefore　establishing　water　security　environment　monitoring　and　evaluation　mechanisms　is　a　good　way　for　providing　services　and　solutions　for　regional　environmental　management　and　accident　pollution.　Considering　features　of　super-large　water　transmission　engineering,　water　environmental　safety　level　assessment　strategy　based　on　intelligent　video　surveillance　is　designed.　According　to　features　of　main　water　surface　and　that　of　moving　target,　and　by　using　SVM　the　corresponding　water　environment　security　level　classification　model　is　designed.　The　application　of　sample　offline　learning　methods　guaranteed　efficiency　and　accuracy　of　water　quality　assessment,　and　the　results　of　several　kernels　are　also　analyzed　and　verified　via　tests.　This　algorithm　has　been　successfully　used　in　the　central　line　project　of　south-to-north　water　diversion,　and　the　test　result　shows　that　this　algorithm　is　of　high　robustness　and　accuracy.