Online scheduling in batch processes is a difficult problem because of the complexities, dynamic behavior and process variability involved. In this paper, a process-variability-based online scheduling system (PVSS), which modifies a schedule based on the current process state and possible future processing time variations at short intervals and at the completion of each task, is proposed. The objective of schedule modification by PVSS is to maintain the performance of the schedule and to control the waiting time of each task within the range specified in a recipe under processing time variations. The effectiveness of PVSS is demonstrated by simulation experiments on a multiproduct batch process model. The results indicate that online scheduling systems should modify schedules based on the possible process variability in the future to improve their performance under processing time variations.