thanks Michael and Erik,

actualy I read the same pulse for every run, the difference is under 10 pulse for each run and up to 10 pulse difference is ok for my system, or maybe up to 50 step.
because the mechanical construction is well designed and constructed thats why I was not thinking to closed loop system, but in real found that after couple of hours the stepper motors start to miss some steps even that I used a good brand.
I think there is a conceptual mistake in my work, is it true that for each step on the encoder there should be a quadrature pulse on A and B and Z phase, regardless of their width.