Water transparency is an important ecological indicator for shallow lakes. The largest shallow lake, Poyang Lake, as well as the most typical river-connected lake in China was selected as the research area. In view of the complicated water-sediment conditions induced by its frequent water exchange with external rivers, the dominant factors driving water transparency were determined against the field investigated data from 2003 to 2013 and a specific driving function was established. A numerical model coupling suspended sediment, Chl-a and chemical oxygen demand was developed and validated, and the spatial water transparency distributions under three typical current structures in Poyang Lake, Gravity-style, Jacking-style and Backflow-style, were quantitatively estimated. The following results stood out: water transparency in the lake varied distinctly with the current status; Backflow-style current was basically characterized by the lowest water transparency, while that under Jacking-style was the highest due to the lower sediment carrying capacity. In some outlying regions in the lake, where the water current is hardly influenced by the mainstream, the water transparency was always kept at a stable level.