In　view　of　the　difficulty　of　capacity　measurement　and　low　accuracy　in　predicting　remaining　life　of　lithium-ion　battery,　a　predictionmethod　of　the　remaining　life　of　lithium-ion　battery　based　on　discharge　process　is　proposed.　Firstly,　the　discharge　time　for　a　specific　voltage　interval　of　lithium-ion　battery　is　extracted　as　the　health　factor.　Then,　bidirectional　extreme　learning　machine　is　constructed　as　the　prediction　model　of　battery　remaining　life.　Finally,　the　accuracy　of　the　prediction　method　is　verified　based　on　the　experimental　data　of　lithium-ion　battery.　The　results　show　that　the　model　can　accurately　predict　the　remaining　life　of　battery,　and　has　smaller　error　and　faster　convergence　speed　compared　with　the　common　extreme　learning　machine　model　and　BP　neural　network　model.　©　2021,　Society　of　Automotive　Engineers　of　China.　All　right　reserved.