Based on the existing geomagnetic diurnal variation theory and correction method, this study comprehensively analyzed the international open geomagnetic diurnal variation data by Fourier transform and one-dimensional continuous wavelet transform, and judged that the frequency band greater than 0.2 Hz is the embodiment of the magnetic disturbance change field in the frequency domain, and used one-dimensional discrete wavelet transform to separate the data, and thus improved the existing geomagnetic diurnal variation correction method. According to the characteristics of daily variation data and one-dimensional discrete wavelet decomposition, this study selected the Sym8 wavelet as the basic wavelet to decompose the data at 7 scales, and the long-term and short-term changes of daily variation data are effectively separated from the magnetic disturbance part under the condition of ensuring the fidelity. Compared to the Fourier series decomposition and low-pass filtering, the processing effect of one-dimensional discrete wavelet transform got better results. The quality of daily variation correction in marine geomagnetic survey can be improved by using the effective separation and correction of short-term, long-term changes and magnetic disturbances in daily variation data, which reduce the error accumulation in the process of marine geomagnetic data processing, and improve the scientificity and accuracy of the current diurnal variation correction methods.