Extreme　events　with　high　impact　and　low　probability　pose　a　serious　threat　to　the　safe　operation　of　power　and　transportation　networks.　To　effectively　enhance　the　resilience　of　power-transportation　coupled　networks,　this　paper　proposes　a　scheduling　optimization　strategy　for　multi-type　mobile　emergency　resources　based　on　a　hybrid　data-model　driven　approach.　Firstly,　considering　the　influence　of　power　line　outages,　traffic　road　faults　and　dynamic　changes　of　traffic　flow　distribution　on　the　strategies　of　mobile　energy　storage　system　(MESS),　line　repair　crew　(LRC)　and　road　repair　crew　(RRC),　a　stochastic　optimization　scheduling　model　of　multi-type　mobile　emergency　resources　to　minimize　the　total　loss　cost　of　the　power-transportation　coupled　networks　is　constructed.　Secondly,　this　paper　proposes　a　hybrid　data-model　driven　approach　to　solve　the　complex　nonlinear　stochastic　optimization　model.　In　the　data-driven　part,　a　novel　graph　diffusion　attention　network　multi-agent　reinforcement　learning　algorithm　is　proposed　to　solve　the　mobile　emergency　resources　routing　strategy.　In　the　model-driven　part,　this　paper　converts　the　power　system　model　and　the　multi-type　mobile　emergency　resources　scheduling　model　into　a　mixed　integer　second-order　cone　programming　model　to　solve　the　mobile　emergency　resources　optimal　scheduling　strategy.　Finally,　the　effectiveness　of　the　proposed　method　is　validated　in　the　modified　IEEE　33-bus　power　network　and　12-bus　traffic　network.　©　2025　IEEE.