The traditional regulation of district heating (DH) system are mainly based on the central control in heat plants and/or in the substations. But the end-user control is becoming increasingly important in saving DH energy consumption. In view of this, a novel zoned group control method of indoor room temperature based on model predictive control (MPC) for a space heating building is proposed. Firstly, a room temperature simulation model based on IDA Indoor Climate and Energy (IDA-ICE) is developed for the experimental room, where the simulation result is validated by the measurements. But it is still difficult to use this model in the predictive control. Therefore, a resistor capacitor (RC) room temperature simulation model with parameters identified by IDA-ICE is proposed. A room temperature controller based on RC-MPC is developed and tested in the experimental room for model validation. Secondly, a DH building in Dalian, China is chosen for the case study. The IDA-ICE simulation results show that the temperature distribution in different rooms of the building is clearly uneven without end-user control. Corresponding targeted control strategies according to the temperature distribution should be used in different zones (cluster of rooms) to improve indoor thermal environment. Therefore, a zoned group control model based on RC-MPC for DH building room temperature is developed in MATLAB/Simulink, which adopts IDA-ICE for parameter identification in RC simulation model and utilizes RC model to predict the room temperature trajectory in the MPC controller. Results show that the proposed model can achieve active and dynamic room temperature control for different zones more precisely according to the changing heat load distribution. It is not necessary to define every room a group, instead, many rooms with similar conditions can be grouped in the same cluster, and temperature differences between rooms can be controlled within a reasonable range.