The　short-circuit　faults　are　one　of　the　inevitable　faults　types　in　the　power　system.　The　frame　circuit　breaker　can　quickly　break　the　transmission　lines　in　case　of　the　short-circuit　faults　to　achieve　the　purpose　of　protecting　the　power　system.　At　present,　the　method　for　detecting　the　short-circuit　faults　in　the　low-voltage　distribution　system　by　the　intelligent　release　in　the　frame　circuit　breaker　is　mostly　the　rate　of　change　of　current　or　the　effective　value　of　current.　Detection　of　the　short-circuit　faults　by　the　rate　of　change　of　current　has　poor　anti-interference　ability.　Detection　of　the　short-circuit　faults　by　the　effective　value　will　take　a　long　time.　In　order　to　shorten　the　short-circuit　fault　detection　time　of　the　intelligent　release,　this　paper　proposes　the　wavelet　packet　decomposition　as　the　short-circuit　fault　detection　algorithm　of　the　intelligent　release.　The　multi-scale　wavelet　algorithm　has　a　certain　ability　to　filter　white　noise,　and　the　singular　value　of　the　detail　decomposition　can　be　used　to　fast　detection　of　the　short-circuit　faults　when　the　short　circuit　occurs,　which　can　shorten　the　detection　time　of　the　short-circuit　faults.　The　results　of　the　authoritative　institutional　tests　prove　that　the　intelligent　release　designed　in　this　paper　can　detect　the　fault　characteristics　and　send　out　the　tripping　signal　within　0.5　ms　after　the　short-circuit　faults　occur.　The　method　proposed　in　this　paper　can　be　used　in　the　low-voltage　distribution　system,　as　well　as　in　medium-voltage　and　high-voltage　situations　to　detect　the　short-circuit　faults.