Principal Component Analysis (PCA) is well known for its effectiveness in dimensionality reduction and denoising. However, the presence of missing data in real-world scenarios often presents a formidable challenge to its application. Practitioners are usually faced with the choice of either imputing the missing data before implementing PCA on the restored dataset or attempting to extract the principal components directly from the incomplete dataset without any prior imputation. In this paper, we present a novel approach known as PCA for missing data (PCAM). This innovative technique is specifically designed to derive principal components directly from missing value datasets, reducing the need for preliminary imputation. Experimental results indicate that PCAM showcases competitive performance when juxtaposed with state-of-the-art methods from both imputation-first and direct extraction strategies, thereby demonstrating its potential to revolutionize the approach to handling missing data in PCA.