I am having a multispectral sensor with Red band, Blue band, Green band, Red-edge band and Near-Infrared band. I need to find water stress using these bands, is it possible to find and if yes what will be index to calculate ? I am using UAV for this task.
Depending on the camera you are using, and what bands, (Sounds like the micasense), you may want something that reaches towards the SWIR to see water characteristics.
Anyways, from what I know- it is best to pair thermal imagery (measuring canopy temperature) with atmospheric temperature, and use something like the crop water stress index (CWSI) - see paper by Bellvert et al. 2014 as an example.
With the typical VNIR (Visible to Near Infrared) indices you will be able to detect vegetation stress... whether that stress is associated with water, or something else entirely, will have to be verified by ground truthing. There are no VNIR indices for discretely detecting water stress (at least that I know about).
I advise you to do more reading of academic journals. Another good paper is by Calderon et al (2014) : Detection of downy mildew of opiumpoppy using high-resolution multi-spectral and thermal imagery acquired with an unmanned aerial vehicle.
What you ask now is a tricky question that many Agro-consulting companies are still researching. This is cutting edge stuff. Consider your thermal resolution, and thermal sensor's spatial resolution as well. A lot of factors can influence and contaminate temperature at the canopy level - as noted by many researchers - thus it is a hard experiment to control. Read up and make a sound methodology.
As for veg stress - there are countless combinations of indices asides from NDVI which can potentially detect stress. This indicates that you should do more reading and literature review. Note that I have experience with micasense, and such a sensor has trouble isolating stress by type (fungal, water, nutrients) in cereal crops. Thus, perhaps thermal will aid in your research.