The target goal of this work is to develop a methodology of forecasting Latvian GDP using ARMA (AutoRegressive-Moving-Average) and VAR methods. The paper follows up with the papers published in the proceedings of the APLIMAT journal in 2011 – see [1]. The algorithm is developed for finding optimal time series model for GDP forecasting. Latvian GDP data with quarterly observation frequency is taken as time series. ARMA and VAR Analysis of Latvian GDP, M2X and inflation indicators time series is performed. The set of models has been constructed. In order to check the accuracy of models, different residual tests are performed: autocorrelation, Portmaneteau, heteroscedasticity and normality of residual distribution. Models are compared in their forecast quality.