In aquatic environments, organisms are exposed to trace metals through a number of routes, including water, sediment, suspended particles and food chain.In this study, the bioavailability and toxicity of several metals in water and sediment was investigated using carefully designed bioassays and modeling methods.　　First, the biouptake of the lanthanide Sm in a freshwater green alga Chlamydomonas reinhardtii was studied within the framework of the biotic ligand model (BLM).The uptake of Sm3+ in the absence of organic ligands was well described by a Michaelis-Menten equation, suggesting a single-site transport.The addition of several simple organic ligands decreased Sm influx rates, however to much less extent than predicted, presumably due to the direct internalization of the Sm complexes.The competition effects of major cations and three other lanthanide cations were successfully modeled by the BLM.The stability constants and Jmax were very similar among the four investigated lanthanides, suggesting a common uptake pathway among them.　　Second, the deposit-feeding sipunculan worm, Phascolosoma arcuatum, was used as a tool to assess metal bioavailability in marine sediments.The worm was considered a suitable test organism because it indiscriminately ingests sediment particles and has very low uptake rates of dissolved metals.In addition, the worm has simple anatomy and is like a little sac full of liquid, i.e., coelomic fluid, which can be easily collected for metal analysis after simple sample treatment.Analyzing the metals in coelomic fluid led to similar results as the somatic-tissue metals for assessing the bioavailability of sediment-bound metals and the spatial pattern of metal contamination.　　Third, a two-compartment toxicokinetic-toxicodynamic model was developed for metals in a freshwater model organism Daphnia magna.Model parameters were estimated for three metals, i.e., cadmium, zinc and mercury, by fitting the literature data on metal bioaccumulation and toxicity.A range of crucial information for toxicity prediction can be readily derived from the model, including detoxification rate, no-effect concentration, threshold influx rate for toxicity, maximum duration without toxicity.This process-based model is flexible and can help for improving the ecological risk assessments for metals.