FPM is better for moving mesh....Especially for moving/deforming domain, adaptive refinement and multiscale phenomena...FPM is suggested to yield high accuracy in point results 

Thank you for your answer.

May i interpret your reply such that fpm is superior to fdm/fem/fvm on solving PDEs in terms of accuracy. If so, Will you please tell me any resource or research paper to justify the same.

Meshless methods are especially useful for domains that change shape substantially in time (e.g. free surface flows in which the free surface is not a single-valued function). In simpler situations, it is probably just a matter of taste whether you want to use a mesh or not. In many applications a mesh is necessary anyway to define external forcings, so it makes sense (to me at least!) to use it also for approximation purposes. In fluid flow problems over time independent domains, some meshless, purely lagrangian methods have the problem that the 'particles' employed tend to follow the flow, so that some areas of the domain may end up to be not sufficiently well resolved. This can be cured by adding more 'particles', but this entails some arbitrariness and possibly other numerical issues. They can be overcome, but this motivates some people to stay with mesh based methods unless your domain is really strongly time dependent.