Sketch person re-identification aims to identify images with identities similar to those of sketched person images located in an RGB image gallery. Although several cross-modal retrieval algorithms can be adopted for this purpose, the background settings of such algorithms are relatively simple and fail to consider that certain identities have only one modal sample in the training set, that is, the cross-modal identity is inconsistent. This significantly limits the application of such algorithms in practical scenarios. In this paper, a sketch re-identification network based on cross-classification is proposed. The network consists of two parts: cross-classification and identity information alignment based on distance. Among these, cross-classification guides the encoder to extract modal-invariant information from one modal using constraints of the classifier trained using other modal data. The alignment of identity information based on distance can reduce the feature distance between different modals of the same identity, suppress the influence of cross-modal identity inconsistencies, and strengthen the discrimination and robustness of features. To verify the performance of the re-identification network when the cross-modal identity is inconsistent, a new sketch re-identification dataset is generated based on Market-1501. The Rank-1 is improved by 11.0 percentage points on this dataset. Simultaneously, the model also achieves a Rank-1 of 60% on the public dataset Sketch Re-ID. The dataset used in this study is an open-source dataset available on “https://github.com/huangdaichui/Sketch_dataset”.