In the last ten years, the increasing power of computers and graphics cards has stimulated developers and users to deepen Virtual Reality (Huang & Lin, 1999; Kreuseler, 2000; Bernhardt et al., 2002). One of its natural applications of interest for the academic and the industrial communities is the Terrain Visualization Systems (TVS) (Lin et al., 1999; Luebke et al., 2003; Ellul & Haklay, 2006). In this field, two-dimensional representations have been completely superseded since three-dimensional (3D) visualization is closer to reality as well as easier to interpret. Furthermore, it allows the user to interact realistically with the environment (Hirtz et al., 1999; Kersting & Döllner, 2002). Nowadays, there exist lots of issues to bear in mind on designing a TVS: various Web3D technologies to represent the Digital Terrain Elevation Models (DTEM), several ways to provide the TVS with sufficient realism, many graphic libraries both private and opensource projects, different applications to enable the user to interact with the system, etc. For example, some recent works can been reviewed about the use of non-proprietary available Web3D technologies (Web3D Consortium, 2006), specifically VRML, X3D and Java3D (Fairbairn & Parsley, 1997; Huang & Lin, 2002; Hay, 2003; Geroimenko & Chen, 2005; Hirtz et al., 2006). The creation of a DTEM can be referred for instance in (Ayeni, 1982; Longley et al., 2001; Fencík et al., 2005). There exist also several databases that can be used as starting point (GTOPO30, 2006; ETOPO2v2, 2006; Gittings, 1996). It is necessary to operate with these databases to extract a grid according to the necessities of realism and fast screen refresh required for this type of application. Furthermore, a correct structure and nomenclature for this grid must be carried out, in order to facilitate and to expedite its management. Another important aspect is the inclusion of texture-mapping in the model to give realism to the visualization (Heckberts, 1986; Guedes et al., 1997; Döllner et al., 2000). The applied textures are the terrain orthophotographs, which are previously treated -to readjust them according to the coordinates of the VRML environment-, partitioned -with the appropriate size for the different elements that constitute the DTEM- or properly structured in order to improve the interactive visualization of massive textured terrain datasets if needed. Regarding the VRML viewers that can be used to represent the DTEM, some well-known samples are DeepViewTM , CASUSPresenterTM , WorldViewTM , BS Contact, Viewpoint Media Player, Emma 3D (Emma3D, 2006; g3DGMV, 2006; Universal 3D Format, 2006). Once 3D navigation system is developed, some interaction tools can be added using VRML Script and Java (Moore et al., 1999). As it can be seen, there are several developments all over the world and very recent semantics in 3D visualization, so it is necessary to make an special effort in generating surveys and standards (Duke et al., 2005). In this chapter, we wish to contribute to clarify the process of development of a TVS in real time by providing a guide through the issues previously commented and illustrating the stages with practical examples. We explain the pros and cons of some of the different currently available options, offering criteria for an appropriate development. We come out of our experience in the Renewable Energy Research Group of La Rioja (Spain) to illustrate this guide. In order to overcome the limitations given by Web3D technologies in general, and VRML in particular, a specific graphic engine developed with open source graphic libraries is shown. Some programs - used to rename the terrain textures according to general VRML structures - and small applets, as interaction tools between the user and the 3D scene, have been implemented in a virtual TVS of La Rioja (one of the 17 autonomous regions in Spain, with a surface of about 5000 Km2). They are used to clarify and exemplify some issues throughout the chapter. The chapter is organized as follows. First, the basic characteristics of a TVS will be briefly commented in Section 2. The Web3D-VRML technologies are introduced in Section 3 where their strong and weak points are shown. Section 4 is devoted to explore the VRML viewers and some tips to create the DTEM and to endow the TVS with interactivity are provided. The development of the graphic engine, and its libraries, which present the 3D geometry of the scene, are discussed in Section 5. Finally, Section 6 concludes the chapter and refers to future work.