Objective: Nosocomial sepsis is among the major factors contributing to mortality in intensive care units (ICUs). Mortality predictors in general ICU patients with nosocomial sepsis were investigated. Methods: This retrospective cohort study was conducted between January 1, 2013 and May 1, 2014 in two general ICUs of a training and research hospital. In total, 95 sepsis attacks developing in 83 patients were included in the study. Data from patients' medical records were recorded on standardised forms. Results: Sepsis was detected in 21.2 cases per 100 ICU admissions. The median length of ICU stay was 37.56±39.595 (range, 1-173) days. Study population consisted of 43 (51.8%) male and 40 (48.2%) female patients. Their ages ranged from 18 to 90 (mean, 69±15.753) years. The median APACHE II score was 26. 9±6.4 (range, 15-45). The primary reasons for admission were medical problems in 62 (74.7%), elective surgeries in 13 (15.7%), and emergency surgeries in 10 (12.8%) patients. Pneumonia (80%) accounted for the majority of nosocomial cases of sepsis detected in the ICUs. Pseudomonas aeruginosa (24.6%), Acinetobacter baumannii (24.6%), and Klebsiella pneumoniae (18.5%) were the most frequently isolated microorganisms. Rate of mortality secondary to nosocomial sepsis was 41 percent. In conclusion, multivariate logistic regression showed that emergency surgery (p=0.004), an increase in the SOFA score (p=0.001), and haemodialysis required for acute renal failure (p=0.004) were statistically significant risk factors for mortality due to nosocomial sepsis. Conclusions: Monitoring SOFA scores may be useful for the monitorization of the patients with nosocomial sepsis.