Placing of concrete requires much more water than the cement needs for its hydration. This results in a certain porosity in the hardened concrete, which facilitates chemical degradation of this material. By adding small amounts of polymeric admixtures, called superplasticizers, one can greatly decrease the amount of water required to obtain the desired workability and consequently the porosity, which improves durability. Although these superplasticizers are widely used today, the mechanisms through which they act in concrete remain poorly understood. It nevertheless recognised that the origin of the effect, which superplasticizers have in concrete, comes from decreasing the attractive forces between cement particles. The effective volume of agglomerates and thereby the effective volume of solids in the suspension is decreased, which improves workability. Controling and modifying interparticle forces has been for many years, and remains today, a topic of intense research in the field of colloids. Also, is it not surprising that many authors have attempted to apply this knowledge to explain the effect that superplasticizers have in concrete. They have thereby been able to identify different types of behaviour depending on the chemical nature of superplasticizers. However, applying such concepts developed for colloids, to cement suspensions in which particle sizes are much larger presents many limitations. The main objective of this thesis has been to integrate these limitations, as rigorously as possible, while applying interparticle force concepts to cement suspensions. In order to reach this objective, further development of current theoretical approaches had to be carried out. In particular, it was necessary to integrate the non-ideality of the thermodynamics of the aqueous phase of the cement suspension. These effects were integrated into the calculation of electrostatic and dispersion forces. In addition, we have taken into account the very large particle size distribution of cement, by determining the frequency of contacts between particles of different sizes. This allowed us to develop a quantitative relation between calculated interparticle forces and measured yield stress of cement suspension. Although this approach still presents limitations, it is the first approach, at least for cement, which gives more than a qualitative link between rheology and interparticle forces. These calculations could only be developed after experimental data had allowed attribution of the dispersing effect of superplasticizers exclusively to the adsorbed polymers. For this, we had to use, in addition to cement, inert model systems (Mg(OH)2 and MgO) in order not to superimpose to desagglomeration, the effect of modifying the on-going chemical reactions. The interest of these model systems over others is that they have a high isoelectric point (around pH 12) like cement. It is therefore possible to have both surface charge and pH similar to cement. This is important for adsorption and polymer conformation to be representative of cement suspensions. We argue that the major effect of chemistry of cement suspensions is the coprecipitation or intercallation of the polymers during the formation of ettringite. This reaction reaches its maximum speed barely a few minutes after water is contacted with cement. It is why one observes much larger polymer consumption, when it is included in the mixing water. This has been often cited in literature, but was never developed in a satisfying way with respect to dispersion efficiency. The calculation of interparticle forces, in a framework adapted to cement, as well as the identification of the reactions consuming superplasticizers, open new paths for the development of superplasticizers, which will offer higher performances for lower dosage. This should make the immediate cost of durable concrete more affordable, which is probably the best incentive to reduce the energy cost of the life service maintenance of concrete infrastructures.