FreqMamba is a novel image deraining method that integrates spatial domain sequence modeling with frequency domain global modeling. The method leverages the complementary strengths of Mamba and frequency analysis to enhance the ability to perceive global degradation and local details. The core of FreqMamba is a three-branch structure: spatial Mamba, frequency band Mamba, and Fourier global modeling. Spatial Mamba processes original image features to extract details and correlations. Frequency band Mamba decomposes the image into sub-bands of different frequencies and performs frequency dimension scanning. Fourier modeling captures global degradation patterns using Fourier transform. The method also incorporates degradation priors derived from rainy images at different scales to generate attention maps, which are integrated into the network to improve training efficiency. Extensive experiments show that FreqMamba outperforms state-of-the-art methods in both visual and quantitative metrics. The method is effective for image deraining and can be extended to other image restoration tasks such as low-light image enhancement and real-world image dehazing. FreqMamba demonstrates its versatility and robustness in various image restoration scenarios.FreqMamba is a novel image deraining method that integrates spatial domain sequence modeling with frequency domain global modeling. The method leverages the complementary strengths of Mamba and frequency analysis to enhance the ability to perceive global degradation and local details. The core of FreqMamba is a three-branch structure: spatial Mamba, frequency band Mamba, and Fourier global modeling. Spatial Mamba processes original image features to extract details and correlations. Frequency band Mamba decomposes the image into sub-bands of different frequencies and performs frequency dimension scanning. Fourier modeling captures global degradation patterns using Fourier transform. The method also incorporates degradation priors derived from rainy images at different scales to generate attention maps, which are integrated into the network to improve training efficiency. Extensive experiments show that FreqMamba outperforms state-of-the-art methods in both visual and quantitative metrics. The method is effective for image deraining and can be extended to other image restoration tasks such as low-light image enhancement and real-world image dehazing. FreqMamba demonstrates its versatility and robustness in various image restoration scenarios.