This article intends to explore the limits of scanning with the technology of 3D Laser Scanner and the 3D printing, as an approximation to its application for the survey and the study of singular elements of the architectural heritage. The case study we developed is the Porta del Drac, in the Pavelló Güell, designed by Antoni Gaudí. We divided the process in two parts, one about how to scan and optimize the survey with the Laser Scanner Technology, made with a Faro Forus3D x330 scanner. The second one, about the optimization of the survey as a highresolution mesh to have a scaled 3D model to be printed in 3D, for the musealization of the Verdaguer House of Literature in Vil.la Joana (Barcelona), a project developed by the Museum of History of Barcelona, in tribute to Jacint Verdaguer. In the first place, we propose a methodology for the survey of this atypical model, which is of special interest for several factors: the geometric complexity in relation to the occlusions, the thickness of the metallic surfaces, the hidden internal structure partially seen from the outside, the produced noise in its interior, and the instrumental errors. These factors make the survey process complex from the data collection, having to perform several scans from different positions to cover the entire sculpture, which has a geometry composed of a variety of folds that cause occlusions. Also, the union of the positions and the average of the surfaces is of great relevance, since the elements of the sculpture are constructed by a metal plate of 2mm, therefore, the error in the union of all these many positions must be smaller than this. Moreover, optimization of the cloud has a great difficulty because of the noise created by the instrumental error as it is a metal sculpture and because of noise point clouds that are generated inside the internal folds of the wings, which are made with a welded wire mesh with little spaces between them. Finally, the added difficulty that there is an internal structure between elements of the parts of the Drac that are partially hidden and therefore cannot be recorded. Secondly, we expose the procedures performed to move from a point cloud to an optimal high-resolution mesh to be printed in 3D, adapting it to all the limitations that this printing technique entails. On the one hand, for the meshing process, a previous classification of the point cloud by pieces (wings, chains, mosaics, head ...) is made and an internal structure is re-assembled to avoid floating parts. On the other hand, the selection of the 3D printing technique, in this case FDM (Fused Deposition Modelling), limits the size of the model so it needs to be cut by determined maximum dimension, and also it limits the minimum thickness of the model’s surface, that is to say, the model cannot be directly scaled to the desired size because the 2mm surfaces would be too thin to be printed. This research intends to advance the knowledge of data acquisition, optimization, modelling and 3D printing, with a case study of great complexity. A process that can be systematized and applied to other models.