To a substantial extent, riparian forests and their diversity depends on soil types and the associated water regime and sometimes water quality. Factors to consider could be numerous, but you might want to stratify habitats into groups such as hydric, non-hydric soils or drainage classes, flood frequency associated with stream type, perhaps geology and valley type, etc. There are ways to study water frequency using wells, stream and rain gauges, interpreting soil indicators of saturation, etc. The US Forest Service has developed various approaches to sample diversity such as publication on Groundwater Dependent Ecosystems as one example, and various publications on riparian ecosystems.

I think along the riverside can layout the sample plot and can carried out the inventory. I think this can simply work to find out the diversity and composition.

Yes, like Sony said, just simple inventories

I think the method must classify the type of soil, the level of humidity, slope and distance of the wood plant from the edge of the river.

I think you should explore the TRARC method that has been used very successfully in Australia and Sri Lanka . https://wetlandinfo.ehp.qld.gov.au/wetlands/resources/tools/assessment-search-tool/15/index.html