Autogenous shrinkage is the unrestrained volume change of cementitious materials occurring at con-stant temperature without any change in mass. It occurs as a consequence of self-desiccation (de-crease of internal relative humidity) and increasing capillary pressure in the pore fluid. In concrete elements hardening in sealed conditions, autogenous shrinkage is critical for crack development, in particular in high performance concrete with low water to cement ratio. Being able to predict autog-enous shrinkage can provide insights for the long-term deformation of different concrete mixtures, create a useful database for simulations and ultimately minimize the cracking risks. The main objective of this thesis is to predict the autogenous shrinkage of fly-ash-blended cement systems by using a quantitative multi-physics approach. In order to reach this goal, a systematic ex-perimental study of autogenous deformation, self-desiccation, microstructure evolution, elastic properties and basic creep of different cementitious systems was first carried out. After having achieved a phenomenological understanding, supported by experimental data, analytical and numer-ical modeling of autogenous deformation is possible. In this study, a novel method based on the evolution of microstructure for predicting the self-desiccation was developed. In this method, 1H nuclear magnetic resonance and mercury intrusion porosimetry were combined to obtain the evolution of the microstructure. This prediction provides another possibility to predict capillary pressure and corresponding autogenous shrinkage of a simu-lated microstructure. Prediction of autogenous shrinkage, including poro-elastic and poro-visco-elastic response based on experimental quantities, was accomplished. In the prediction, the poro-elastic deformation of ce-mentitious materials was calculated based on poromechanics. The poro-visco-elastic response was studied with basic creep tests on hardening cementitious materials. Generalized Kelvin-Voigt chains were applied to predict the aging of creep. The prediction matched reasonably well with measured autogenous shrinkage. Autogenous shrinkage of cementitious materials was also numerically modeled with a microstructure simulation and a finite element method. This part of the work extended existing approaches based mainly on two programing platforms: µic platform and Automatic Mechanics for Integrated Exper-iments (AMIE) finite element framework. Average pore pressure load calculated from measured relative humidity, saturation degree and Biot coefficient was imposed into a 3-dimensional micro-structure computed with µic platform. Based on back-calculated elastic and visco-elastic behavior of C-S-H, an approach for simulating the autogenous shrinkage was demonstrated.