PURPOSE: Carbapenem-resistant Acinetobacter baumannii (CRAB) represents a critical global health threat for which existing antibiotics are increasingly inadequate. This study aimed to establish a comprehensive genomic framework for the rational prioritization of virulent Acinetobacter bacteriophages as therapeutic candidates. METHODS: We performed large-scale comparative genomic analysis of 340 virulent Acinetobacter bacteriophages, integrating phylogenetic reconstruction, pangenome analysis, CRISPR spacer-based host interaction mapping, Anti-CRISPR protein identification, and systematic antimicrobial resistance (AMR) gene screening. RESULTS: Genome sizes spanned a nearly 20-fold range, with a significant negative correlation between genome size and GC content (R² = 0.139, ρ = -0.630). Phylogenetic analysis revealed extensive divergence across multiple lineages with no dominant clade. Pangenome analysis identified 20,982 unique protein families, of which 76.2% were cloud genes, confirming a highly open genome architecture. CRISPR spacer matching yielded 1,480 high-confidence matches across 100 phage genomes, providing molecular evidence of broad historical infectivity. Anti-CRISPR profiling identified Acinetobacter phage XC1 as an exceptional therapeutic candidate harboring 55 predicted Anti-CRISPR proteins with canonical regulatory locus architecture. AMR screening identified 21 distinct AMR gene homologs (Loose RGI hits, 22.5 to 47.1% amino acid identity) distributed heterogeneously across the dataset, confirming abundant therapeutically clean candidates while flagging a subset warranting further scrutiny before therapeutic exclusion. CONCLUSION: These findings provide a multi-criteria genomic framework for rational phage candidate prioritization against multidrug-resistant Acinetobacter infections, with direct implications for evidence-based phage therapy development.