The construction of spatial geographic units is regarded as a prerequisite and foundation for the study of crime distribution. Geographic units, when constructed from the criminal travel perspective are considered conducive to a better understanding of the determinants influencing criminals' choices of crime locations. Current research on the spatiotemporal distribution of crime and its influencing factors is predominantly based on administrative divisions, grids, and roads; however, the spatial characteristics of criminal travel are not addressed in these studies. To address this gap, a new method for constructing geographic units oriented towards criminal travel was proposed, and its impact on the distribution of street thefts was discussed. Firstly, the spatial characteristics of criminal travel were reviewed, and a new method for constructing geographic units based on subway stations and bus stops as reference points was proposed. Secondly, an empirical study was conducted using Chaoyang District in Beijing as a case study. From the criminal's perspective, environmental factors were sorted out, and criminal perception factors affecting travel decisions were identified. A comparative Analysis was performed to assess the impact of perception factors on street theft distribution between the proposed geographic units and traditional administrative division units. Results indicate that the model efficiency, predictive effect, and explanatory level of perception factors using the geographic units are all stronger than those of administrative division units. The proposed geographic units are shown to better capture spatial heterogeneity and the influence of population flow on crime. The geographic units oriented towards criminal travel proposed can better reflect patterns of criminal travel, explain the reasons for crime distribution, and enrich research in the field of criminal geography.