In the end of 1990s it was found that the universe is undergoing an accelerated expansion phase, suggesting the presence of a mysterious component called dark energy, which with dark matter forms the so-called dark sector of the universe, responsible for about 95 % of its content. Gravitational waves can be used to investigate this dark sector, complementing current electromagnetic observables such as type Ia supernova and cosmic microwave background. From the amplitude of the gravitational wave it is possible to measure the luminosity distance d L from the generating source and with the electromagnetic counterpart it is possible to determine the redshift z and thus build a d L − z diagram and measure the cosmic expansion history. Thus, gravitational waves can be used as standard sirens. The first standard siren detected was gravitational waves originating from the collision of a pair of neutron stars on August 17, 2017 (GW170817) where the electromagnetic counterpart (GRB170817A) was also detected. With this event it was possible to determine that the propagation velocity of gravitational waves and the velocity of light differ in one part by 10 15 , which allowed us to restrict in many models that tried to explain the cosmic acceleration through theories of modified gravitation. In this thesis I will present two applications. The first will be about a dark-sector unification model from Horndeski Theory, and how event GW170817 imposed strong constraints on this model. The second will be about simulating standard siren catalogs that could be detected by the future Einstein Telescope and how these future data could constrain dark sector interaction.