This paper focuses on a Bayes inference model for a simple step-stress life test using Type-I censored sample in a discrete set-up. Assuming the failure times at each stress level are geometrically distributed, the Bayes estimation problem of the parameters of interest is investigated in the both of point and interval approaches. To derive the Bayesian point estimators, some various balanced loss functions are used. Furthermore, a simulation study and sensitivity analysis is performed to carry out the performance of the results of the paper. An example is also presented to illustrate the proposed procedure. Finally, some conclusions are stated.