Comparing processes has important application value in business process management, and it is common in scenarios of process version control, and process similarity calculation, process merging. Mapping elements between process models is the first step for process comparison. Current comparison solutions, however, only consider the mapping between tasks but ignore the correspondence of other elements in the process. Hence, the reliability of the results from the process comparison cannot be guaranteed. To address this issue, this paper first proposes a Petri net based model for element mapping between processes, and focuses on description of the context that is required for places mapping. Then, based on the place mapping context, a mapping algorithm is also presented. For any pair of processes modeled by Petri net, the algorithm adopts bilateral and unilateral mapping strategies where the pair of places that shares the most similar contexts is given high priority to be selected as the mapping result. The extensive experiments based on the real world data sets show the effectiveness of the proposed method as well as its efficiency in the scenario of process similarity search.