Lithium-ion　batteries　have　become　a　most　promising　energy　storage　candidate　in　power　station　and　electric　vehicles　because　of　its　high　power　capability,　high　energy-conversion　efficiency,　and　environmental　friendliness.　It　is　significant　to　diagnose　the　security　of　battery　by　monitoring　the　its　state　parameters.　Wherein,　temperature　and　strain　are　the　two　of　the　important　ones.　In　this　work,　a　sensitivity-enhanced　FBG　strain　sensor　was　designed　for　the　strain　measurement　of　lithium-ion　batteries.　This　proposed　sensor　consists　of　two　FBGs　and　a　lever　mechanism.　The　lever　mechanism　works　as　a　displacement　amplifier.　The　amplified　deformation　of　battery　act　on　the　functional　FBG　and　induce　the　larger　wavelength　shift.　The　thermal　compensation　FBG　can　eliminate　the　influence　of　ambient　temperature.　The　calibration　test　shows　that　this　sensor　has　a　high　sensitivity　of　11.55　pm/μ　and　a　good　linearity.　Application　test　on　a　battery　illustrates　that　the　strain　responses　of　the　sensor　has　a　good　repeatability　in　three　cycles.　Then,　artificial　neural　networks　were　used　for　state　of　charge　(SOC)　estimation.　When　the　strain　and　temperature　data　were　set　as　input　parameters,　SOC　can　be　well　predicted.　Therefore,　this　sensor　can　monitor　the　strain　on　the　cell　with　high　sensitivity　and　accuracy.　This　research　demonstrated　a　new　solution　for　SOC　estimation　especially　based　on　strain　signals,　which　can　provide　more　informative　data　for　battery　management　system.　©　2019　SPIE.