Flow pattern recognition in porous media is crucial for debris bed modeling and state tracking. Traditional measurement techniques face challenges due to the complex structure of porous media. Although X-ray computed tomography can provide detailed flow images, its application is constrained to small-scale laboratory settings and entails high costs. This study introduces an intelligent analysis method using Principal Component Analysis (PCA) to identify two-phase flow patterns in porous media. Key features were derived from differential pressure signals corresponding to typical flow patterns (bubbly, slug, and annular flows). These features were further analyzed using probability density function (PDF) and power spectral density (PSD) techniques. Five characteristic parameters were developed to capture the essential features of flow patterns: gas and liquid Reynolds numbers, mean differential pressure, PDF skewness, and PSD kurtosis. A new PCA-based characteristic equation was developed, achieving a 97.67% accuracy rate in identifying typical flow patterns. The PCA model presented in this study enables automated classification of flow patterns in porous media, providing a valuable tool for addressing challenges in debris bed formation and flow characterization.