Concrete is a microheterogeneous material. Therefore, mechanical properties of concrete are related to the hierarchically organized microstructure of the material. At an observation scale of millimeters to centimeters, one can visually distinguish sand grains, gravel aggregates, and the surrounding cement paste matrix. Resolving the cement paste matrix at smaller scales of observation reveals a surprisingly complex material microstructure. It consists of cement grains, pores, and hydrates; whereby the latter represent products of the chemical reaction between cement and water. Material modeling is particularly challenging at early material ages, because the microstructure of cement paste undergoes a continuous transformation due to the progressive consumption of cement and water, and the corresponding precipitation of hydrates. Describing the evolving microstructures within a mathematical framework is the first objective of this work. We aim at quantifying the volumes occupied by the material constituents, as functions of the initial volumetric composition and of the maturity of the material. Thereby, we account for recently quantified phenomena like the progressive densification of calciumsilicate- hydrates (C-S-H) and the -internal curing- capacity provided by water residing in the open surface porosity of aggregates. Additional important challenges tackled in this thesis are: identification of the morphology of the individual material constituents and of their arrangement within the hierarchically organized microstructure, quantification of their mechanical properties, and modeling their interactions. Corresponding multiscale models are fed with measured or modeled input data, taken from several fields of cement science reported in the open literature. The mass density and the elastic stiffness of solid C-S-H nanoparticles are taken from small angle scattering experiments and from atomistic modeling, respectively. Strength properties and the densification behavior of C-S-H gel are taken from limit state analysis of nanoindentation tests and from nuclear magnetic resonance relaxometry tests, respectively. This way, the number of model parameters is kept at an absolute minimum and all involved quantities are physically meaningful. Methods of continuum micromechanics are used as vehicles for scale transitions, i.e. for establishing links between microstructure and microstructural properties, on the one hand, and macroscopic mechanical properties of cementitious materials, on the other hand. Bottomup homogenization is used to upscale physical laws introduced at material microscales and top-down identification is used to quantify constants of material constituents, which are nowadays not accessible by direct material testing. Thereby, the present thesis addresses all three major mechanical properties of cementitious materials: their elastic stiffness, their creep properties, and their uniaxial compressive strength. As for poroelasticity, it is shown that stiffness homogenization starting at nanoscopic solid C-S-H particles all the way up to the macroscopic elastic behavior of cement paste is possible, if one considers that space confinements in the water-filled pore spaces govern (i) the shape of precipitating solid C-S-H particles and (ii) the overall density of the evolving C-S-H gel. As for creep, it is shown that the maturity- and composition-dependent creep properties of cement pastes, mortars, and concretes - as quantified in several thousands of macroscopic creep experiments - can be traced back to one universal creep function of microscopic hydrates. As for strength, it is shown that hydrates of environmentally friendly -green- cement pastes and mortars, produced with slag or fly ash as cement replacement materials, are considerably stronger than the hydrates in ordinary Portland cement pastes.