Carbon-based conductive fillers have been incorporated into cement matrix to develop smart self-sensing materials with piezoresistive properties. However, accurately measuring the sensing property of the cement composite without compromising its mechanical performance is not easy to achieve in practical engineering. Therefore, in this study, a novel experimental setup for measuring the self-sensing properties of conductive fillers embedded cementitious composites was developed. This multi-functional measurement system is able to measure specimens under compressive and flexural stress with different loading profiles, apply various loading rates, obtain the electrical properties, and measure the strain using both LVDT and Particle Image Velocimetry (PIV) or Digital Image Correlation (DIC) techniques with all the data synchronised to one file sharing the same time stamp controlled by Python codes. In this study, the piezoresistivity and the performance on damage detection of the cementitious composites with low graphite concentration (5%) in a bulk form were investigated through monotonic compressive and flexural tests. Experiment results include the specimens’ stress, strain and Fractional Change in Resistivity (FCR). Data analysis showed that the set-up and methodology developed in this study are effective to test self-sensing cementitious composites, and the graphite-cement composites used in this study have a stable piezoresistivity and able to detect damage upon failure.