A　novel　multi-objective　reactive　power　optimization　method　of　distribution　network　was　proposed,　which　used　complex　affine　power　flow　algorithm　to　compute　the　uncertainty　resulted　from　DGs.　It　first　established　the　interval　multi-objective　optimization　model　of　network　loss　and　voltage　deviation.　Then　it　described　the　DGs＇　output　and　state　variables　as　complex　affine　and　computed　the　interval　power　flow　by　Ybus-Gauss　iteration.　In　the　process　of　optimum　seeking　with　NSGA-II　algorithm,　it　divided　the　individuals　into　several　groups　according　to　dominance　relationship　and　calculated　the　interval　crowded　distance　in　the　non-dominated　layer　to　eliminate　the　faultage　of　Pareto-optimal　front.　The　validation　of　the　proposed　methodology　was　demonstrated　by　using　an　IEEE33　system　with　DGs.　The　results　shows　that　the　proposed　method　can　optimize　the　network　loss　and　voltage　simultaneously　as　considering　the　uncertainty　of　DGs.　©　2017　Chin.　Soc.　for　Elec.　Eng.