A　combination　weighting　approach　is　proposed　to　solve　the　weight　coefficients　of　the　objective　function　for　the　selection　of　DC-link　location,　which　adopts　the　entropy　function　to　consider　both　the　randomness　of　weights　and　the　consistency　among　weight　vectors,　constructs　a　constrained　nonlinear　model　and　applies　the　improved　particle　swarm　optimization　algorithm　to　solve　it,　resulting　in　a　more　reasonable　weight　combination　and　a　more　reliable　objective　function.　Comparison　of　the　average　difference　between　the　proposed　and　other　objective　functions　shows　that,　the　proposed　approach　for　the　selection　of　DC-link　location　is　effective　and　more　reasonable.　©,　2015,　Electric　Power　Automation　Equipment　Press.　All　right　reserved.