二类水体主要包括内陆及近岸水体，受浮游植物、悬浮颗粒、有色可溶性有机物等多种因素影响，光学特性复杂多变，难以建立统一的水环境参数遥感定量估算模型。针对水体的光学特征，进行水体光学分类，进而反演水环境参数的方法，不仅能够提高参数估算精度，而且便于模型在同类水体中推广应用。水体光学分类方法主要包括基于固有光学特征的光学分类、基于遥感反射率波形特征的光学分类和以参数反演为目标的光学分类等方法。在分类反演的策略中，包括分类与模型算法融合、基于水体光学类型优选算法、优选多模型混合计算等方法。具体应用时，需要根据研究区水体光学特征的复杂程度和研究目的，选取不同的分类方法及参数遥感估算策略。

Case II waters, including inland and inshore waters, are affected by many factors, such as phytoplankton, suspended particles, and colored dissolved organic matter, leading to complex and changeable optical characteristics of the water body. Hence, establishing a unified remote sensing quantitative estimation model for retrieving water environmental parameters is difficult. According to the optical characteristics of water, the method of water optical classification and water environmental parameter inversion can not only improve the accuracy of parameter estimation but also facilitate the model to be popularized in similar waters. This study aims to review the state-of-the-art concepts and methods of water optical classification on remote sensing technology for case II water monitoring. The classification-based applications on retrieving environmental parameters as well as the limitations and prospects are discussed.The criteria for considering studies for this review are based on the general development of water optical classification technology and ongoing studies from authors and their collaborators. The selection of studies is classified by different methods and applications for parameter retrieval. The main concept and advantages of water optical classification are illustrated with several examples presented in this study.According to the optical characteristics of water, the method of water optical classification and water environmental parameter inversion can not only improve the accuracy of parameter estimation but also facilitate the model to be popularized in similar waters. Water optical classification methods mainly include optical classification based on inherent optical characteristics, remote sensing reflectance waveform characteristics, and parameter inversion. The classification inversion strategy includes the fusion of classification and model algorithm, the optimization algorithm based on water optical type, and the hybrid calculation based on optimization of multi-model.Water optical classification is an effective tool for remotely recording the water quality and improving the estimation of the parameters especially in optically complex case II waters. The water retrieval of one predominated optical type should be based on its optimal model. However, accurate estimation of water composed of various types with spatiotemporal dynamics requires the determination of optimal models for each type and the blending strategy. The fuzzy-logic-based blending supports the production of seamless contiguities by considering weight factors. However, different classification methods and parameter estimation strategies should be reconsidered according to the complexity of water optical characteristics and research purposes.