Why most only one variable change in an experiment?

That is simply not true.

There are times when a number of factors can have an effect on the outcome of an experiment but it is not possible to conduct the experiment so that only one factor is allowed to vary while all others are controlled. It is far more efficient - in terms of experimental units - to design the experiment so that the interactions between variable factors can be estimated. Experimental design is the subject that deals with the problem of designing efficient schema for dealing with a number of interacting factors. You may wish to look at "Latin square design" and "Greco-Latin squares" as an introduction to design.