Caridean shrimps are the second most diverse group of Decapoda. Over the years, several different systematic classifications, exclusively based on morphology, have been proposed, but the classification of the infraorder Caridea remains unresolved. In this study, five nuclear genes, 18S rRNA, enolase, histone 3, phosphoenolpyruvate carboxykinase and sodium–potassium ATPase α-subunit, were used to examine the systematic status of caridean families and superfamilies. We constructed gene trees based on a combined dataset of 3819bp, containing 35 caridean species from 19 families in 11 superfamilies. At the family level, and based on our restricted representation, our molecular data support monophyly of the families Glyphocrangonidae, Crangonidae, Pandalidae, Alpheidae, Rhynchocinetidae, Nematocarcinidae, Pasiphaeidae, Atyidae and Stylodactylidae. In contrast, both the Hippolytidae and Palaemonidae are polyphyletic in our analysis. Two major clades are revealed. The Alpheidae, Hippolytidae, Crangonidae, Glyphocrangonidae, Barbouriidae, Pandalidae, Hymenoceridae, Gnathophyllidae and Palaemonidae make up the first clade, while the second clade comprises the Rhynchocinetidae, Oplophoridae, Nematocarcinidae, Alvinocarididae, Campylonotidae, Pasiphaeidae and Eugonatonotidae. Two families, Bathypalaemonellidae and Stylodactylidae, are shown to be basal groups in our tree. At the superfamily level, our results do not support the currently accepted superfamily classification, although there is support for a superfamily Palaemonoidea, though only three out of its eight families are included. The results suggest that the currently accepted superfamily classification of the Caridea does not reflect their evolutionary relationships. A major revision of the higher systematics of Caridea appears thus to be vital, ideally incorporating both molecular and morphological evidence.