Background: Colistin is increasingly used to treat severe infections caused by multi-drug-resistant (MDR) bacteria, particularly in critically ill patients. Its effectiveness, especially in monotherapy, remains controversial. This study aimed to evaluate the effectiveness and toxicity of colistin therapy in severe MDR infections. Methods: This retrospective study included patients treated with colistin (CMS) at the ICU. Patients’ treatments were divided into four subgroups: monotherapy vs. combination therapy, empirical vs. targeted therapy, intravenous vs. intravenous plus inhaled therapy, and standard doses with and without a loading dose. The primary outcome was clinical cure. Secondary outcomes included microbiological eradication, survival rate, and drug-related toxicity, particularly acute kidney injury (AKI). Exclusion criteria included Gram-positive infection, inhaled therapy alone, use of colistin <5 days. Results: A total of 150 patients (mean age 60 ± 18 years, APACHE II score 17 ± 10) were included. The most frequent condition was hospital-acquired pneumonia (n = 140, 93.3%). The most common pathogen was MDR Acinetobacter baumannii (n = 146, 97.3%). In most patients, colistin therapy was targeted (n = 113, 75.3%) and combined with other antibiotics (n = 124, 82.7%). Inhaled CMS was added in 47 (31.3%) patients. Mean duration of therapy was 10 ± 4 days. Clinical cure occurred in 64 (42.7%) patients, microbiological eradication in 20 (13.3%). AKI developed in 65 (53.7%) patients. Inhaled CMS improved the clinical cure rates (57.4% vs. 37.0%, p = 0.003). Conclusions: Intravenous CMS was mainly used for MDR Acinetobacter baumannii-related pneumonia. Clinical cure was observed in 42.7% of patients, but renal toxicity was high. Combining intravenous and inhaled CMS may improve outcomes.