This article discusses the feasibility of immersed visible / near-infrared spectroscopy for rapidly detecting the content of haematococcus pluvialis chlorophyll. Combined with chemometrics methods, we chose the best pretreatment to predict the content of haematococcus pluvialis chlorophyll a and chlorophyll b. We compared the results of PLS modeling of full-wave band and the successive projections algorithm combined with PLS (SPA-PLS) modeling, and chose the better one for prediction. For chlorophyll a and b, SPA-PLS modeling was better than PLS modeling of full-wave band. The RPD value of chlorophyll a by SPA-PLS modeling reached 2.946 1 and that of chlorophyll b by SPA-PLS modeling was 1.902 3. Research showes that this method can predict the chlorophyll a, chlorophyll b content well through some pretreatments algorithms and modeling algorithms which can provide a new approach for detecting the content of chlorophyll a and chlorophyll b.