Generally the driest month is the best for differentiating between land uses, end of the dry season if there are a number of dry months. Make sure if you are comparing between years that you check rainfall data, you don't want to be comparing dry and wet season. 

The first condition is that the month you are considering should be cloud-free and further your specific objective. 

The original post has already acknowledged accounting for atmospheric effects (clouds). That said, I think it's necessary to be more specific about the study area. India is a diverse country, covering many latitudes, ranging from tropical conditions to the Himalayas. In an area with diverse vegetation, you might need multiple months (i.e. leaf-on and leaf-off) to do an adequate classification. In areas without deciduous vegetation or major seasonal crop variations, one month might suffice. In other words, it's really impossible to generalize the answer for the entire country of India.

It depend on the satellite image. find the suitable image which is captured in the day which there is low cloud cover approximatly less than 9 %.

Hi Gh,

as you are analysing land use, it may be useful to first determin the main abiotic limiting factor/s to the success of the landuse intent.  Then correspond the seasonal time period in which the abiotic stressor occurs; this will help target a period for investigation.

all the best

pam

Generally the images used for land use changes have cloud cover less than 5%

yes otherwise cloud will cover your area and will not get good output(cloud less than 10 %)