In　order　to　achieve　rapid　and　non-destructive　detection　of　the　astaxanthin　content　in　Haematococcus　pluvialis,　a　snapshot　multispectral　imaging　method　was　proposed　in　this　paper.　An　imaging　system　was　built　by　using　two　snapshot　multispectral　cameras　with　visible　spectral　ranges　of　480~635　nm　and　near-infrared　spectral　ranges　of　665~950　nm,　respectively.　The　spectral　data　of　H.　pluvialis　samples　in　different　growth　periods　was　collected.　To　optimize　the　model　prediction,　a　great　variety　of　methods　were　compared,　including　different　spectral　ranges,　three　preprocessing　methods,　two　characteristic　wavelength　selection　methods　and　two　modeling　methods.　The　results　indicated　that　the　combination　of　both　visible　and　near-infrared　spectroscopy　achieved　the　optimial　prediction　performance　with　the　pretreatment　of　first　derivation　(FD),　and　the　characteristic　band　selection　method　of　competitive　adaptive　reweighting　sampling　(CARS)　and　modeling　method　of　back　propagation　(BP)　neural　network,　the　prediction　set　correlation　coefficient　(Rp)　of　0.9622,　the　root　mean　square　error　(RMSEP)　of　0.5126　and　the　residual　prediction　error　(RPD)　of　3.6726,　which　was　superior　to　the　visible　alone　(Rp　of　0.9467,　RMSEP　of　0.6065　and　RPD　of　3.1042).　These　indicated　that　it　was　feasible　to　detect　the　content　of　astaxanthin　in　H.　pluvialis　by　the　snapshot　multispectral　imaging　technique,　and　the　combination　of　both　visible　and　near-infrared　spectroscopy　could　be　more　effective.　©　2023　Editorial　Department　of　Science　and　Technology　of　Food　Science.　All　rights　reserved.