The short-term scheduling problem (STSP) of oil refinery operations with unreliable distillation units (CDUs) is a NP-hard problem, necessitating a high degree of accuracy and efficient techniques. Developing effective scheduling optimization techniques for oil refineries is a significant challenge. In order to provide solutions to this difficult problem, this study proposes a new hybrid Petri net and tabu search (PNTS) approach with unreliable CDUs to effectively solve this problem. To show the efficiency of the developed approach, an industrial case study is used. The results demonstrate that the developed method efficiently solves the industrial case study.