Fever is one of the most common symptoms in children. However, although the frequency of fever as the main symptom is high, there is considerable controversy in the management of children that have fever without source. The aim of this study was to determine the incremental rate of cost-effectiveness of presepsin compared with C – reactive protein (CRP) and procalcitonin in detecting severe bacterial infection (SBI) in children aged 29 days to 36 months that have fever without source according to pre-test probability in Colombia. An economic analysis of cost-effectiveness structured using the technique of decision analysis (TreeAge Pro ®) was performed. Systematic reviews and meta-analysis were analyzed to establish the sensibility and specificity of CRP, procalcitonin and presepsin, a consensus of experts determined the length of hospitalization, costs of hospitalization and tests evaluated were determined from the average direct costs (USD). The cutoffs used were < 0.5 mg/dl for procalcitonin, > 40 mg/L for CRP and 625 pg/ml for Presepsin. The pre-test probability ranged from 10% to 90%, considering low, intermediate and high probability of SBI. The different strategies had similar cost-effectiveness for a correctly diagnosed patient with SBI. However, presepsin was the most C/E strategy for the pretest probability scenarios between 30% - 90%, ranging from $911 to $2685 per diagnosed patient. In the lowest pre-test probability, 10%, the CRP performed better. In the clinical practice, a large amount of children with fever without source have a wide range of pre-test probabilities of SBI. Our results found that presepsin can be a good diagnostic tool in patients with a 30% or higher probability of presenting SBI in children in Colombia. Additional research of new diagnostic tools is necessary to improve care evidence in children with SBI.