[en] In this thesis we investigate various aspects of flux compactifications in six-dimensional quantum field theories. After introducing the internal geometries, i.e. the two-dimensional torus T"2 and one of its orbifolds T"2/Z_2, we classify possible gauge backgrounds including continuous and discrete Wilson lines with emphasis on a non-vanishing flux density. An operator analogy with the quantum harmonic oscillator allows for an explicit derivation of the mode functions of charged fields and demonstrates the advantage of our interpretation of discrete Wilson lines in terms of localized fractional gauge fluxes. We then derive a globally supersymmetric action which captures the D-term supersymmetry breaking induced by the internal magnetic field and reproduces the Landau level mass spectrum of the charged four-dimensional degrees of freedom. In this context we show that, even though supersymmetry is broken at the compactification scale, the inclusion of the whole tower of charged states leads to vanishing quantum corrections for the Wilson line effective potential on T"2. This result is supported by a symmetry breaking argument in which the Wilson line appears as a Goldstone boson. After that, we additionally include gravitational effects within a supergravity effective action of the lightest modes in four dimensions. The dynamics of the moduli fields arising after compactification can be encoded in the setup of N=1 supergravity augmented with anomaly cancellation by the Green-Schwarz mechanism. This leads to a non-trivial transformation behavior for two axion fields under gauge variations in the low-energy effective action. As an application, we discuss an SO(10) x U(1) grand unified theory which uses the multiplicity of fermionic zero modes in the flux background to induce the number of matter generations. Finally, we investigate a novel mechanism for generating de Sitter vacua in N=1 supergravity based on a flux-induced positive definite D-term potential. The inclusion of a superpotential, leading to negative contributions in the energy density, allows a successful stabilization of all fields. Employing a general approach to decouple the Goldstino from the rest of the fermions, we evaluate the masses of all light degrees of freedom explicitly.