The　load　increase　or　DG　(Distributed　Generator)　integration　of　distribution　network　may　advance　or　delay　the　expansion　of　power-supply　equipment　and　then　cause　the　change　of　capacity　cost.　The　sensitivity　of　branch　current　to　the　nodal　injection　power　is　deduced　according　to　the　power　flow　of　radial　distribution　network,　which　is　combined　with　the　node-branch　power-flow　correlation　matrix　to　build　the　corresponding　sensitivity　matrix　for　calculating　the　LMCC　(Locational　Marginal　Capacity　Cost)　of　distribution　network.　For　the　DG　owner,　LMCC　can　be　used　to　evaluate　the　DG　capacity　benefit　under　different　electricity　price　systems;　for　the　grid　with　DGs,　it　can　be　used　to　build　a　multi-type　DG　planning　model,　which　is　then　solved　by　the　particle　swarm　optimization　algorithm.　Case　study　verifies　the　feasibility　of　the　proposed　model　and　method.　©　2016,　Electric　Power　Automation　Equipment　Press.　All　right　reserved.