In the geomagnetic data processing, crucial for applications such as MT data processing, and critical mineral exploration, the need to fill data gaps in geomagnetic records, extract solar quiet (Sq) signals even on days of heightened solar activity, and discern disturbance components is widely acknowledged. Presently, there is no suitable method to effectively undertake these tasks. I propose a novel approach termed EigenMag to address this requirement by leveraging Principal Component Analysis (PCA). The initial phase of EigenMag involves the transformation of geomagnetic records into a feature space, followed by the conversion of these features into the eigen-vector space through PCA. Subsequently, EigenMag constructs gaps in records, Sq, and disturbance signal components in the eigen-vector space. Validation of EigenMag was conducted using synthetic gaps generated from geomagnetic records, wherein the constructed gaps were compared with the original signals. The results indicate that EigenMag performs exceptionally well, particularly during periods of low and moderate solar activity. Additionally, a comparison between the separated components with the global disturbance indicator index (ap) revealed a strong correspondence. Further analysis involved a comparison of the fractal properties of raw geomagnetic records with their corresponding parts (loading scores) in the eigen-vector space. This comparison indicated that geomagnetic signals are suitable for gap filling in the eigen-vector space than in the temporal space. In conclusion, EigenMag has been successfully employed for tasks such as gap filling and the construction of Sq and disturbance components, showcasing its efficacy in enhancing geomagnetic data processing.