<p>Single-Cell Clustering (SCC) is a specialized algorithm developed for analyzing single-cell data. It utilizes unsupervised methods to infer cell identity labels from gene expression profiles, and has emerged as a rapidly advancing field of research. SCC continues to play a critical role in various biomedical downstream applications, including differential expression and pseudotime analysis. Current SCC models can generally be classified into three categories based on the type of encoder: Embedded, Topological, and Multi-Modal. Early research primarily concentrated on traditional clustering algorithms and single-modal cell analysis, while recent advancements have expanded these models to incorporate multimodal approaches and deep neural networks to capture nonlinear relationships. However, comprehensive reviews and discussions of this critical area are currently lacking in existing research literature and open-source repositories. To fill this gap, we have conducted an extensive review of SCC algorithms. Specifically, we propose a dual-level classification framework that organizes the approaches based on the encoder architectures driven by biological structural priors (Embedded for feature-level, Topological for relation-level, and Multi-Modal for modality-level structures) and the timeline (Traditional and Deep clustering). Furthermore, we examined publicly available datasets in the field and assessed the strengths and limitations of current methods, identifying existing challenges and potential avenues for future research.</p>