Bipolar disorder (BD) and schizophrenia (SCZ) are serious psychiatric conditions that share similar clinical characteristics and genetic predispositions. Nevertheless, the fundamental molecular mechanisms linking these disorders have not been fully explored. This research seeks to clarify the common molecular pathways and potential therapeutic targets in BD and SCZ by employing an integrated bioinformatics and machine learning methodology. Gene expression datasets were sourced from public databases, and an analysis of differential gene expression was performed using the Limma package. A feature selection method based on filtering, specifically Fisher Score, was utilized to pinpoint the most significant genes common to both disorders. We detected 30 shared differentially expressed genes (DEGs) in BD and SCZ. Functional enrichment analysis through Gene Ontology (GO) demonstrated a significant connection to mitochondrial activity and cellular respiration. KEGG pathway analysis additionally indicated links to oxidative phosphorylation and pathways associated with neurodegenerative diseases. The development and clustering of the protein-protein interaction (PPI) network revealed five pivotal hub genes-NDUFS2, NDUFB2, UQCRQ, COX7C, and NDUFA13-all of which are integral components of the mitochondrial respiratory chain, implying a shared metabolic dysfunction in both conditions. Furthermore, regulatory network analysis revealed complex interactions involving transcription factors (TFs), microRNAs (miRNAs), and the common DEGs, highlighting intricate gene regulatory mechanisms. Predictive drug analysis pinpointed potential therapeutic agents aimed at these common molecular signatures. In summary, our results offer new insights into the shared biological foundations of BD and SCZ, providing encouraging pathways for future diagnostic and therapeutic approaches.