Snake venom metalloproteases (SVMPs) comprise zinc-dependent enzymes members of the Reprolysin subfamily, which include the ADAMs (A Disintegrin And Metalloproteinase) domains. These metalloprotesases (MPs) are responsible for a relevant pathophysiology in envenomation, including local and systemic hemorrhage. Alteration of basal membrane components and interactions with some factors of the coagulation pathway are the main targets of these classes of enzymes. P-III class SVMPs, that have a MP, a disintegrin or disintegrin-like and a cysteine-rich domain, are potent hemorrhagic toxins isolated from snake venoms. Important correlations have been established between hemorrhage and these structural domains. BjussuMP-I, a MP from Bothrops jararacussu snake venom, characterized as a P-III class of SVMP, presents 546 amino acids residues, comprising high percentages of polar groups (Asp, Glu, Gly). This protein exhibit proteolytic activity on fibrinogen, acting as an inhibitor of platelet aggregation. In this work we report the initial crystallization trials for BjussuMP-I, and a theoretical model for its catalytic domain. Crystals of BjussuMP-I were obtained at 298K using 0.5M ammonium phosphate. These crystals were submitted to X-ray diffraction studies and diffracted at low resolution. We are now trying to improve the quality of the crystals aiming to get high resolution X-ray diffraction data. A theoretical model for the catalytic domain of BjussuMP-I was built through the folding recognition (threading) techniques using the program Fugue and this model was refined by a molecular dynamics (MD) simulation using the program GROMACS. This model was created through the average combination of three different template structures initially selected in the PDB database. The overall stereochemical quality of the final theoretical BjussuMP-I catalytic domain model was checked through the programs PROCHECK and ProSa2003. The Ramachandran plot showed that 98.4% of the model residues are in the core and additionally allowed regions. Furthermore, the graph calculated by the program ProSa2003 shows that all residues have negative potential energies, which indicates the high stability degree of the model. Thus, all these results demonstrate that BjussuMP-I catalytic domain theoretical model is reliable. The presence of large hydrophobic areas and some conserved surface charge-positive residues were identified as important features of the SVMPs and other metalloproteases.