Thermochemical heat storage is a promising technology towards efficient use of renewable energy resources. Materials based on salts and their hydrates have a high potential for a good energy storage density and the benefit of long-term storage ability. However, the process has not yet been successfully implemented due to limitations in mass and heat transfer. This paper investigates how to improve the less desirable properties of CaCl2 and its hydrates such as low melting points, agglomeration, low cycle stability and low sorption rates. The optimization of CaCl2 properties was achieved by mixing with KCl and impregnation in carrier materials to obtain a composite material. The tests show at first that, with the admixtures of KCl, water uptake during hydration is 2 times higher than that of CaCl2. Water release during dehydration is 1.3 times higher than that of CaCl2. Secondly, the use of compacted expanded natural graphite (ENG) or activated carbon foam (ACF) increases the cycle stability, thermal conductivity and the water sorption performance. Due to their hydrophobic nature those matrices have no influence on the reaction scheme, thus the total amount of water molecules sorbed by the salt-in-matrix is close to the value of CaCl2. The degree of impregnation varies from 31 to 90 wt% depending on the host matrix and the impregnating medium used. The water vapour uptake is up to 0.61 g g−1 and the water released ranges from 0.12 to 0.72 g g−1. The thermal conductivity of CaCl2-in-matrixis is 3 times higher than that of sole CaCl2.