This dissertation focuses on (1) predicting the presence and absence of trout (Salmo trutta) for the whole Iran and (2) on developing fish-based multi-metric assessment indices for cold-water and cyprinid streams for the Caspian Sea Basin. For predicting the presence and absence of trout, 1090 sites were investigated. Five different modelling techniques (GLM, GAM, GBM, CTA, RF) which were summarised in an ensemble forecasting approach were used to explore the potential distribution of brown trout. For developing multi-metric fish index in cold-water streams, 88 sites (44 reference and 44 impacted), and for cyprinid streams 102 sites (50 reference and 52 impacted) were investigated. First,various human pressure types related to land use, connectivity, hydrology, morphology, water quality and biology as well as their combinations were investigated in each stream type. According to the results, most sites were affected by land use pressure in both stream types. In addition, only a few sites were affected by single pressure in both stream types while most sites in cold-water streams were affected by double and triple pressures and in cyprinid streams by multiple pressures. Moreover, the differences of fish metrics between reference and impacted sites were quantified for each stream type according to a site-specific modelling approach. In fact, specific statistical models were used to describe metric responses to natural environmental differences in the absence of any human pressures. By including impacted sites, the residual distributions of these models described the response range of each metric to human pressures, independently from natural environmental influence. Finally, to develop a multi-metric fish index of cold-water streams, two fish metrics were selected which showed the best ability to discriminate between impacted and reference sites. In contrast, for the development of a multi-metric fish index for cyprinid streams, seven metrics were considered.