Generalised additive models (GAMs) have proven to be a useful tool for modelling egg distribution and egg abundance of small pelagic species. This methodology was used in this work to analyse the bluefin (Thunnus thynnus) larval distributions and their interannual changes in the Balearic Sea, between 2001 and 2005. Larval abundance data were collected through the Mediterranean bluefin spawning season during five surveys, specifically conducted for larval bluefin sampling. Different spatial (i.e. latitude, longitude, depth), hydrographic (i.e. temperature, salinity, oxygen, geostrophic velocity) and biological covariates (i.e. zooplankton biomass) were used in the analysis and model selection. For every sampling year, results provided: (1) larval abundance estimates, (2) accurate mapping of larval distribution over the whole surveyed area, and (3) quantitative estimates of the relationships between larval distribution/abundance and the spatial/environmental variables included in the models. This revealed which of these relationships were significant and their spatial and temporal variability