A GPS baseline solution model is presented, based on the Empirical Mode Decomposition (EMD), which has the advantage of eliminating the error effects outside the model. The EMD technique is a new signal processing method for non-linear time series, which decomposes a time series into a finite and often small number of Intrinsic Mode Functions (IMFs). The decomposition procedure is adaptive and data-driven which is suitable for non-linear data series analysis. A multi-scale decomposition and reconstruction architecture is defined on the basis of the EMD theory and the error mitigation model is demonstrated as well. A standard of the scale selection for the elimination of errors, outside the model, was given in terms of the mean of the accumulated standardized modes. Thereafter, the scheme of the GPS baseline solution based on the EMD is suggested. The float solution residuals of the Double-Difference (DD) observation equation are used to extract the errors outside the model applied to modify the GPS DD measurements. Then the float solution was given again and the fixed solution was obtained by a Lambda algorithm. Three schemes are designed to test the proposed model and the experimental results show that the proposed model dramatically improves the reliability of ambiguity resolution after the elimination of errors outside the model.