The aim of this study lies on the empirical application of the tree bagging methodology, in order to predict the insolvency of Portuguese Small and Medium-sized Enterprises (SME) in the agro-industrial sector, one year in advance. The database consists of financial indicators of 243 companies, available at SABI (Iberian Balance Analysis System), all from agro-industrial sector. The proposed model reveals a robust result when compared with traditional parametric models. The results show that two indicators – “short-term liquidity” and “capacity to generate results appropriate to the size” – were the most statistically relevant, both in the Proposed Model and the Logistic Regression model.