That depends on what you want to study, what question you want to ask. You can study kaons at zero chemical potential, as well as at finite chemical potential. Chemical potential is simply a way to impose a certain constraint. If you're interested in studying the strong interactions at finite density, for example, then you work at finite chemical potential. If not, you work at zero chemical potential (for the model at hand in both cases). This is exactly like for any thermodynamical system: when do you work in the grand canonical ensemble (finite chemical potential), when in the canonical or microcanonical ensemble (zero chemical potential).

@stam Nicolis; if I consider a system of hadrons in thermal and chemical equillibrium . Suppose the system of hadrons contain pions, kaons, lambdas, sima etc particles. Now I want to find the number density or pressure of these particles at some finite chemical potential. I can use fermionic or bosonic distribution function(equillibrium distrubution function). Since this equillibrium distribuion function contain chemical potentialalso. When I find number density of pions(bosons) I take mu=0(chemical potential). If I find number density of lambda( I use mu=mus+2*muq) and a fermionic function; because lambda is a fermion. Now If I try to find the number density of kaons(which is also a boson) should I take mu=0 or to take mu=muq-mus..............

please explain

An example of a system with a zero chemical potential is the QGP (quark gluon plasma) created in the central rapidity region in heavy ion collisions at high energy (see ALICE experiment results for example). Out of this soup of quarks and gluons (reaching thermal equilibrium) a large amount of particles are created but always with an equal number of  particles and anti-particles and even of light nuclei and anti-nuclei. See some nice slides like that  http://folk.uio.no/farido/fys3510/HeavyIons2012-Trine.pdf