Background: Giardia duodenalis is a species complex consisting of multiple genetically distinct assemblages. The species imposes a major public health crisis on developing countries. However, the molecular diversity, transmission dynamics and risk factors of the species in these countries are indeterminate. This study was conducted to determine the molecular epidemiology of G. duodenalis infection in asymptomatic individuals in Southern Ethiopia. Methods: From March to June 2014, fresh stool samples were collected from 590 randomly selected individuals. Socio-demographic data were gathered using a pre-tested structured questionnaire. The genotyping was done using triosephosphate isomerase gene-based nested polymerase chain reaction and DNA sequencing. The genetic identity and relatedness of isolates were determined using the basic local alignment search tool and phylogenetic analysis. Risk factors associated with G. duodenalis infection were analysed using binary and multinomial logistic regression models. Results: The results showed that 18.1% (92/509) of the study subjects were infected by G. duodenalis. Among the isolates, 35.9% (33/92) and 21.7% (20/92) were sub-typed into assemblages A and B, respectively, whereas 42.4% (39/92) showed mixed infections of A and B. Most of the assemblage A isolates (94%,31/33) were 100% identical to sequences registered in GenBank, of which the majority belonged to sub-assemblage AII. However, the high genetic variability and frequency of double peaks made sub-genotyping of assemblage B more problematic and only 20% (4/20) of the isolates matched 100% with the sequences. The risk factors of age (P = 0.032) and type of drinking water source (P = 0.003) both showed a significant association with the occurrence G. duodenalis infection. Conclusions: This study established the endemicity of G. duodenalis in Southern Ethiopia. Infection with assemblage A was more frequent than with assemblage B, and the rate of infection was higher in children and in municipal/tap and open spring water consumers than the other groups. Sub-typing of assemblage B and determining the origin of double peaks were challenging. The present study confirms the need for further inclusive studies to be conducted focusing on sub-types of assemblage B and the origin of heterogeneity. [ABSTRACT FROM AUTHOR]