Background: Osteoporosis and sarcopenia are age-related degenerative diseases that frequently co-occur in the older adults, yet their shared molecular mechanisms remain poorly understood. This study aimed to identify common biomarkers. Methods: By utilizing the GEO database, differential and enrichment analyses were conducted to identify genes that are commonly expressed in both diseases. Subsequently, key genes were identified through four types of machine learning and six types of immune infiltration analyses. Finally, the Dual Condition Sarcopenia and Osteoporosis (DSO) model was established by combining bilateral ovariectomy with natural aging to validate the key genes. Results: A total of 577 and 625 genes were identified in the osteoporosis (GSE156508) and sarcopenia (GSE226151), respectively. GO, KEGG, and GSEA analyses revealed that osteoporosis-related genes were enriched in muscle development and myogenesis pathways, while sarcopenia-related genes were linked to osteoclast differentiation and inflammatory responses. Venn analysis identified 20 genes shared by both diseases. Four machine learning algorithms identified five key genes: APOC1, ENPP5, FBXL22, IRS1, and PAQR4. Among them, PAQR4 showed the highest predictive value in ROC analysis and was consistently downregulated in both training and validation datasets. Immune infiltration analysis indicated altered neutrophil levels, notably reduced in osteoporotic skeletal muscle. Experimental validation confirmed decreased PAQR4 expression in both bone and muscle tissues of the DSO model. Conclusions: These findings highlight PAQR4 as a potential biomarker and a candidate molecule of interest for osteosarcopenia.