ABSTRACT. This paper aims to define the determinants of regional growth in Portugal, at the NUTS III level, in a time span between 1999 and 2010. A panel data approach is used as well as the fixed, random and pooled effects models, all of them with and without trend. The results from the panel data and the Hausman test show that the model that best portrays the reality under study is the panel data with random effects model. The performed analysis allowed us to confirm that employment, sectorial GVA, electricity consumption, number of museums and landline phone accesses have a positive association with the regional per capita GDP. Surprisingly, the number of residents, population density, number of medical doctors and technical progress presents a negative correlation instead.