For　the　drawbacks　of　complex　buried　pipeline　structural　performance　function　and　large　amount　of　calculation　of　nonlinear　problems　and　to　obtain　the　influence　on　pipeline　failure　probability　of　the　shear　wave　velocity,　characteristic　period,　pipeline　diameter,　wall　thickness,　seismic　wave　incidence　angle　in　different　site　types.　Firstly　based　on　the　FOSM　of　traditional　reliability　analysis,　the　artificial　neural　network　is　introduced　to　do　intelligent　analysis　and　the　failure　probability　is　calculated　separately　in　different　site　types.　The　comparison　with　FOSM　shows　that　the　results　are　reliable.　Secondly,　the　commonly　used　gray　cast　iron　pipeline　at　home　is　analyzed.　Finally,　some　meaningful　conclusions　that　siteIand　IV　are　respectively　the　most　favorable　and　disadvantaged　sections,　the　buried　pipeline　seismic　capacity　can　be　improved　by　the　method　of　increasing　pipe　diameter　and　wall　thickness　in　site　II,　III,　are　obtained.　©　2015　ejge.