Accurate and timely diagnostics are critical for managing bacterial infections. The current gold standard, culture-based diagnostics, can provide clinicians with comprehensive diagnostic information including bacterial identity and antimicrobial susceptibility, but they often require several days of turnaround time, which leads to compromised clinical outcome and promotes the spread of antibiotic resistance. Nucleic acid amplification tests such as PCR have significantly accelerated the detection of specific bacteria but generally lack the capacities for broad-based bacterial identification or antimicrobial susceptibility testing. Here, we report an integrated assay based on PCR and high-resolution melt (HRM) for rapid diagnosis for bacterial infections. In our assay, we measure bacterial growth in the presence or absence of certain antibiotics with real-time quantitative PCR or digital PCR to determine antimicrobial susceptibility. In addition, we use HRM and a machine learning algorithm to identify bacterial species based on melt-curve profiles of the 16S rRNA gene in an automated fashion. As a demonstration, we correctly identified the bacterial species and their antimicrobial susceptibility profiles for multiple unknown samples in blinded tests within ∼6.5 h.