PyDMD is a Python package that implements dynamic mode decomposition (DMD) and its variants, providing tools for analyzing spatiotemporal data. The package includes advanced methods for handling noisy, multiscale, parameterized, high-dimensional, and nonlinear data. It offers a user-friendly interface, extensive documentation, and a suite of Jupyter Notebook tutorials. The package supports various DMD algorithms, including optimized DMD, bagging-optimized DMD (BOP-DMD), coherent spatiotemporal scale separation (CoSTS), parametric DMD, randomized DMD, and physics-informed DMD. These methods are designed to improve the robustness and accuracy of DMD in real-world applications. The package also includes tools for data preprocessing, time-delay embedding, and visualization. PyDMD is modular, with each DMD variant implemented as a separate module. The package is open-source, well-documented, and continuously updated to incorporate new DMD methods and features. The paper provides an overview of the package's features, the theory behind DMD, practical usage tips, and examples of its application to synthetic data. It also discusses the importance of selecting the appropriate DMD method based on the nature of the data and the problem at hand. The paper concludes by emphasizing the utility of PyDMD as a practical tool for data analysis and a central resource for DMD methods.PyDMD is a Python package that implements dynamic mode decomposition (DMD) and its variants, providing tools for analyzing spatiotemporal data. The package includes advanced methods for handling noisy, multiscale, parameterized, high-dimensional, and nonlinear data. It offers a user-friendly interface, extensive documentation, and a suite of Jupyter Notebook tutorials. The package supports various DMD algorithms, including optimized DMD, bagging-optimized DMD (BOP-DMD), coherent spatiotemporal scale separation (CoSTS), parametric DMD, randomized DMD, and physics-informed DMD. These methods are designed to improve the robustness and accuracy of DMD in real-world applications. The package also includes tools for data preprocessing, time-delay embedding, and visualization. PyDMD is modular, with each DMD variant implemented as a separate module. The package is open-source, well-documented, and continuously updated to incorporate new DMD methods and features. The paper provides an overview of the package's features, the theory behind DMD, practical usage tips, and examples of its application to synthetic data. It also discusses the importance of selecting the appropriate DMD method based on the nature of the data and the problem at hand. The paper concludes by emphasizing the utility of PyDMD as a practical tool for data analysis and a central resource for DMD methods.