Aiming　at　the　problem　that　the　existing　models　can’t　extract　the　hidden　spatiotemporal　features　from　the　data　effectively;　this　paper　proposes　a　multi-feature　fusion　neural　network-combined　model　for　expressway　traffic　speed　prediction.　The　prediction　model　includes　three　modules:　data　preprocessing　module,　spatiotemporal　feature　mining　module,　and　testing　module.　Compared　with　ARIMA,　LSTM,　and　CNN_LSTM　models,　the　combined　model　achieves　higher　speed　prediction　accuracy　by　using　the　data　of　ETC　gantry　system　between　Fuzhou　and　Xiamen　of　Fujian　province.　It　can　provide　effective　reference　for　decision-making,　analysis,　and　dispatch　of　expressway　management　system.　©　2022,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.