3D Virtual Environments (3D VE) are more and more used in different applications such as CAD, games, or teleoperation. Due to the improvement of smartphones hardware performance, 3D applications were also introduced to mobile devices. In addition, smartphones provide new computing capabilities far beyond the traditional voice communication. They are permitted by the variety of built-in sensors and the internet connectivity. In consequence, interesting 3D applications can be designed by enabling the device capabilities to interact in a 3D VE. Due to the fact that smartphones have small and flat screens and that a 3D VE is wide and dense with a large number of targets of various sizes, mobile devices present some constraints in interacting on the 3D VE like: the environment density, the depth of targets and the occlusion. The selection task faces these three problems to select a target. In addition, the selection task can be decomposed into three subtasks: Navigation, Pointing and Validation. In consequence, researchers in 3D virtual environment have developed new techniques and metaphors for 3D interaction to improve 3D application usability on mobile devices, to support the selection task and to face the problems or factors affecting selection performance. In light of these considerations, this thesis exposes a state of the art of the existing selection techniques in 3D VE and the selection techniques on smartphones. It exposes the selection techniques in 3D VE structured around the selection subtasks: navigation, pointing and validation. Moreover, it describes disambiguation techniques providing the selection of a target from a set of pre-selected objects. Afterward, it exposes some interaction techniques described in literature and designed for implementation on Smartphone. These techniques are divided into two groups: techniques performing two-dimensional selection tasks on smartphones, and techniques performing three-dimensional selection tasks on smartphones. Finally, we expose techniques that used the smartphone as an input device. Then, we will discuss the problematic of selecting in 3D VE displayed on a Smartphone. It exposes the three identified selection problems: the environment density, the depth of targets and the occlusion. Afterward, it establishes the enhancement offered by each existing technique in solving the selection problems. It analysis the assets proposed by different techniques, the way they eliminates the problems, their advantages and their inconvenient. Furthermore, it illustrates the classification of the selection techniques for 3D VE according to the three discussed problems (density, depth and occlusion) affecting the selection performance in a dense 3D VE. Except for video games, the use of 3D virtual environment (3D VE) on Smartphone has not yet been popularized. This is due to the lack of interaction techniques to interact with a dense 3D VE composed of many objects close to each other and displayed on a small and flat screen and the selection problems to display the 3D VE on a small screen rather on a large screen. Accordingly, this thesis focuses on defining and describing the fruit of this study: DichotoZoom interaction technique. It compares and evaluates the proposed technique to the Circulation technique, suggested by the literature. The comparative analysis shows the effectiveness of DichotoZoom technique compared to its counterpart. Then, DichotoZoom was evaluated in different modalities of interaction available on Smartphones. It reports on the performance of the proposed selection technique based on the following four interaction modalities: using physical buttons, using graphical buttons, using gestural interactions via touchscreen or moving the device itself. Finally, this thesis lists our contributions to the field of 3D interaction techniques used in a dense 3D virtual environment displayed on small screens and proposes some future works.