Background Microbiological characteristics of sepsis and antimicrobial resistance are well studied, although in State University of Campinas, no data has been published yet. Methods The main agents related to sepsis and antimicrobial resistance were analyzed. The blood culture records requested from 4,793 hospitalized patients were analyzed. The samples were processed using the Bact/Alert® system for agent identification and antimicrobial susceptibility. Results A total of 1,017 patients met the inclusion criteria for a sepsis diagnosis, with 2,309 samples tested (2.27 samples/patient). There were 489 positive samples (21% positive) isolated from 337 patients (33.13%), but more rigorous criteria excluding potential contaminants resulted in analysis being restricted to 266 patients (315 agents). The prevalent microorganisms were coagulase negative Staphylococcus (CNS) (15.87%), Escherichia coli (13.0%), Staphylococcus aureus (11.7%), Klebsiella pneumoniae (9.8%), Enterobacter sp (9.5%), Acinetobacter baumannii (9.2%), Pseudomonas aeruginosa (5.7%) and Candida sp (5.1%). Examining antimicrobial resistance in the agents revealed that 51% of the S. aureus isolates were methicillin-resistant S. aureus (MRSA) and 80% of the CNS isolates were oxacillin-resistant. For A. baumannii, the ideal profile drugs were ampicillin sulbactam and piperacillin/tazobactam, and for P. aeruginosa, they were piperacillin/tazobactam and ceftazidime. Enterobacteria showed on average 32.5% and 35.7% resistance to beta-lactams and ciprofloxacin, respectively. When all Gram-negative bacteria were considered, the resistance to beta-lactams rose to 40.5%, and the resistance to ciprofloxacin rose to 42.3%. Conclusions Eighty percent of the agents identified in blood cultures from patients with sepsis belonged to a group of eight different agents. For empirical treatment, carbapenems and vancomycin unfortunately still remain the best therapeutic choice, except for A. baumannii and P. aeruginosa, for which piperacillin/tazobactan is the best option.