With　the　increasing　penetration　of　distributed　energy　resources　or　loads　managed　by　aggregators　in　a　distribution　network,　a　critical　discrepancy　emerges　between　bidding　power　and　actual　delivered　power　when　grid　constraints　are　ignored.　This　will　bring　deviation　penalties　and　affect　the　safe　and　stable　operation　of　a　distribution　system.　To　bridge　this　gap,　this　study　proposes　a　bi-level　bidding　framework　for　aggregators　that　integrates　distribution　locational　marginal　pricing　(DLMP)　and　safe　operation　power　meeting　the　safety　operation　criteria　of　a　distribution　network.　The　maximization　of　revenue　for　aggregators　is　considered　in　the　upper-level　model,　whereas　with　the　safety　of　power　flow　and　voltage　limits　taken　into　account,　the　lower-level　model　is　developed　to　maximize　social　benefits.　To　ensure　compliance　with　safety　standards　in　the　aggregators＇　bidding　plans,　stage　one　of　the　lower-level　model　is　employed　to　determine　safe　operation　power　and　the　DLMP　obtained　through　stage　two　jointly　guides　aggregators　in　adjusting　bidding　plans.　To　ensure　that　safe　operation　power　constraints　align　with　aggregators＇　bidding　plans　and　maximize　their　profits,　the　model　is　solved　via　the　alternating　direction　method　of　multipliers.　This　algorithm　facilitates　iterative　optimization　of　aggregators＇　bidding　plans,　adjusting　them　until　the　stage-one　safety　constraints　are　fully　satisfied.　This　process　effectively　mitigates　network　congestion　and　voltage　overrun　issues,　enabling　the　DLMP　to　converge　toward　the　predicted　day-ahead　market-clearing　price　that　balances　economic　efficiency　and　operational　security.　Finally,　a　modified　IEEE33-bus　system　is　used　to　verify　the　rationality　and　effectiveness　of　the　proposed　method.