The building sector and its heating and cooling represent one of the major consumer of energy worldwide. Simultaneously, the share of fluctuating generation of renewable energies in the energy mix increases. Therefore storage and demand side management technologies are required. The new adaptive and predictive control algorithm for thermally activated building systems(TABS) based on multiple linear regression(AMLR) presented in this paper enables the application of demand side management(DSM) strategies. Based on simulations, different strategies have been compared with each other. By applying the AMLR algorithm, electricity energy cost savings of 38% could be achieved compared to the conventional control strategy for TABS, while increasing the thermal comfort. At the same time, thermal energy demand can be reduced in the range between 4% to 8%, and pump operation time from 86% to 89%. The building sector and its heating and cooling represent one of the major consumer of energy worldwide. Simultaneously, the share of fluctuating generation of renewable energies in the energy mix increases. Therefore, storage and demand side management technologies are required. The new adaptive and predictive control algorithm for thermally activated building systems (TABS) based on multiple linear regression (AMLR) presented in this paper enables the application of demand side management (DSM) strategies. Based on simulations, different strategies have been compared with each other. By applying the AMLR algorithm, electricity energy cost savings of 38% could be achieved compared to the conventional control strategy for TABS, while increasing the thermal comfort. At the same time, thermal energy demand can be reduced in the range between 4% to 8%, and pump operation time from 86% to 89%.