Latent thermal energy storages (LTESs) in combination with heat pumps and smart control strategies can maximize the utilization of renewable energy sources for heating and cooling. However, smart energy management with model predictive control (MPC) requires monitoring the total energy stored in the LTES, which is determined by the liquid fraction of the phase change material (PCM). Measuring the liquid fraction is challenging and the diverse liquid-fraction sensing approaches pose a trade-off between accuracy and ease of implementation. The present study aims to quantify the effect of the liquid fraction sensing accuracy on the performance of MPC strategies for heating systems with LTES. For this purpose, a residential heating application with an energy management system is analyzed. The heating system consists of a heat pump, an LTES and a photovoltaic array. The heat pump can be driven by the photovoltaic array and the electric grid. The energy management system uses MPC based on Mixed-Integer Linear Programming. Representative seasonal profiles for the heating load and weather conditions are used as forecasts for the MPC. The performance of the energy management system is assessed in terms of total heating cost for different error values in the estimation of the liquid fraction of the PCM in the LTES. The heating cost is found to proportionally increase with the absolute error in liquid fraction due to reduced utilization of the LTES capacity.