The soil information is great significance to the control strategy of the excavation and loading process of the suction hopper dredger,but the suction hopper dredger cannot directly sense the soil information.We conduct linear visualization and feature analysis of nonlinear coupled dredging data based on real ship data through principal component analysis(PCA),and study the correlation between soil information and working condition information.The support vector machine (SVM) is used to construct the classifier to classify and identify the soil conditions.The results show that the method can effectively identify different working conditions,and realizes the indirect perception of soil information.