There are still many question open to completely understand the structure-activity relationships of G-protein coupled receptors. Issues like the actual mapping of the binding site of different subtypes, as well as the mechanism of activation are poorly understood [1]. Accordingly, further studies on the structure-activity relationships are necessary. In this regard, only a few 3D structures are available from X-ray diffraction studies. Computational studies can complement this information through the construction of reliable 3D models. The goal of the present work is to evaluate the effect of using different ligands to obtain reliable models of the three-dimensional structure of a G-protein coupled receptor using a specific template. Specifically, we have constructed in the present work a three dimensional model of the M3 muscarinic receptor by homology modelling, using the X-ray structure of M2 muscarinic acetylcholine receptor as template and the sequence analyses of muscarinic acetylcholine receptor family. Furthermore, we have studied the effect of the ligand used in the modelling process. For this purpose three models of the receptor were built, including one without ligand and two models with the selective antagonists tiotropium and N-Methylscopolamine, respectively docked into the orthosteric binding site. The constructed models were refined using molecular dynamic calculations to analyze the effect of ligand refinement process and to derive significant conformational information. Based on the analysis of the refined models done through the calculations of RMSD, RMSF, visualization of the structures and comparison between the refined models and the crystal structure of M3 muscarinic receptor, the addition of a ligand in construction of homology model (and the subsequent refinement process) stabilize the structure. Furthermore, the similarities in the structure conformations of both refined models of M3 muscarinic-ligand complexes and the crystal structure of the M3 muscarinic receptors, suggest that the methodology used in this study can be used in prediction of 3D structure prediction GPCR in the absence of crystal structures.