It is really difficult to provide a simple answer to your question. The answer depends a lot on the goals of your work. We know that water quality in a lake can vary spatially and with depth. If you need to characterize this spatial variation, then you must sample the lake at several locations and at multiple depths. The answer also depends a lot on the size of the lake. Small lakes in the range from a few to maybe 100 ha may only need to be sampled in one or a small number of locations. Available funds for water quality characterization will play an important role as well. Here in New York State in the US, we have a lake water quality characterization program in which samples are collected monthly at two depths---one meter below the surface, and from just below the top of the hypolimnion. Samples are always collected in the same location within the lake so that water quality trends over time can be detected. If your lakes show seasonal stratification, then water quality above and below the thermocline can be quite different. At a minimum, seasonal changes are usually tracked, which means at least four times per year in a northern temperate climate. Seasonality may be quite different in your location in India. I hope this information helps, and good look with your work.
Annual water quality assessment of lake or pond how far the temporal span or duration should be there between retrieval of samples - this is your question. For that, I have found one suitable reference in WHO pdf. I attach the pdf here for your reference.
In general I agree with previous comments. However these comments are only valid for monitoring some elements of quality.
In Europe since 2000, with the Water Framework Directive, the concept of water quality has led to a concept of ecological status. Thus, we have to analyze several elements of both biological, physico-chemical and hydromorphological quality. Obviously each requires a different frequency. For example, for phytoplankton Douglas comments seem appropriate, but if we analyze aquatic macrophytes or macroinvertebrates (also required in the WFD), an annual basis is optimal.
Take a look to your year round wheather and then decide, it is good to look at the short and mid term prediction weather models too, because if you only make seasonaly sampling you may be loosing some other important events of lesser time spam.