Hi Sisira - there are a few approaches depending on resources available and water quality data you have and what you want to use it for. This can range from determining filterable metals (more useful than total), to labile measurements (such as Chelex labile - shown to be correlated with ecotoxic effects at least for Cu - see Article A rapid Chelex column method for the determination of metal ...
and other work by these authors - particularly Simon Apte), to chemical speciation models and biotic ligand models (a few developed in N. America and Europe (see http://bio-met.net/)