Background Emerging evidence highlights the dysregulation of mitophagy, the process of clearing damaged mitochondria, as a potential contributor to Alzheimer's disease (AD) pathology. However, the precise mechanisms linking mitophagy to AD remain poorly understood. This study utilized summary-data-based Mendelian Randomization (SMR) combined with multi-omics data to explore the causal relationships between them, and to uncover potential epigenetic mechanisms of gene regulation. Methods Mitophagy-related genes were identified through the integration of three databases, and transcriptomic data of AD patients were obtained from the Gene Expression Omnibus (GEO) database. A meta-analysis was conducted to recognize differentially expressed genes (DEGs) associated with mitophagy in AD. Through SMR tools, genome-wide association study (GWAS) summary data of AD from the GWAS Catalog (n=487,511) were separately integrated with expression quantitative trait loci (eQTLs) and DNA methylation quantitative trait loci (mQTLs) from blood and brain tissues to identify potentially causal genes and methylation sites. The findings from primary analysis were validated with data from the UK Biobank (n=301,478). Results In total, 111 mitophagy-related genes were found to be differentially expressed in AD. The three-step SMR analysis identified two genes, PARL and BCL2L1, from blood tissues and three genes, ATG13, TOMM22, and SPATA33, from brain tissues as causal candidate genes associated with AD. The analysis pinpointed the possible epigenetic mechanisms, where specific methylation sites regulate the expression of these genes, potentially contributing to their association with AD. All findings were successfully replicated in UK Biobank cohorts. Conclusions The study emphasized the putatively causal relationships of mitophagy-related gene with AD. These underlying pathogenic mechanisms could pave the way for new approaches in early detection and therapeutic intervention for AD.