This study attempts to explain the variability in recruitment of sardine in the northern Benguela and to develop potential models by including environmental information to predict recruitment. Two different recruitment and spawner number datasets were available: a VPA-developed dataset, for the period 1952-1987, and data from a simple age-structured model for 1992-2007. In all, four environmental indices were used: the degree of the intrusion of the warm Angola Current into northern Namibia, termed the Angola-Benguela front index; the extent of the upwelling area off central Namibia; average sea surface temperature (SST) over the northern and central Namibian shelf; and wind stress anomalies at Lüderitz as an indicator of upwelling strength. Contrary to general belief, it was found that extremely high recruitment can happen at low spawner levels. This occurred in years in which a large upwelling area existed in association with the minimum southward intrusion of the Angola Current. These effects override the normal negative linear relationships with SST and the positive linear relationship with wind. However, when the area of upwelling is average or small, the effects of spawner biomass, SST and wind become important factors in the variability of recruitment. To estimate exceptional recruitment, the upwelling and front indices were included in the model. To measure medium and weak recruitment, spawner numbers and the SST and wind anomaly formed part of the model. These models can be used simultaneously to predict recruitment before annual acoustic surveys take place and thus aid management decisions.