Invasive Salmonella infection, which can cause typhoid/paratyphoid fever and invasive non-typhoidal salmonellosis, is a public health burden in Africa. Accurate diagnosis and etiological characterization are required to conduct prevalence and risk estimations for Salmonella infection; however, the utilization of optimal techniques and surveillance data are still insufficient. In this study, we performed a laboratory-based survey in Freetown, which is the biggest city in Sierra Leone with a high burden of typhoid fever, by using blood culture and molecular methods but not the Widal test, to estimate the prevalence and aetiology of invasive Salmonella infection among fever patients. We found a very low prevalence of typhoid fever in patients with fever during the investigation period, and this prevalence was clearly overestimated by the Widal test. Genome sequencing of the S . Typhi isolate from this work revealed that the strain carried multiple antibiotic resistance genes, and an epidemic clone that has existed in West Africa for years was also detected in Sierra Leone. By using metagenomic sequencing, one patient with invasive non-typhoidal salmonellosis was identified as having bacterial co-infections. Our data highlight that Salmonella surveillance based on accurate laboratory diagnosis and genome sequencing needs to be strengthened to provide a better estimation of the real epidemics and enable potential risk assessment by etiological analysis in Africa. Even in a laboratory with only basic equipment, it is possible to conduct next-generation sequencing for pathogen discovery in bloodstream infections and to determine the etiological characteristics of pathogene without complex combinations of laboratory methods.