Dear ali,
I assume you are asking about reinforcement bars which are not continuous, but from 2 or more segments connected in some way (typically, there is an overlap where the bar segments run parallel along each other, just bound by some thin wire holding them in place during casting of the concrete)?

Do you expect bond failure in that region to be important for the global response? If not, I would start the modelling with a single continuous bar, neglecting the possibly weak splice connections, and checking the bar stresses in postprocessing to make sure the assumptions for this simplification are valid.

Regards.