Low-voltage flexible interconnection forms interconnection and mutual power supply between low-voltage substation areas in the same regional distribution network, which can improve the total supply capacity (TSC) of the distribution system. Moreover, the distributed energy connected to the low-voltage substation area has temporal and uncertain characteristics, which affect the variation of TSC value. Therefore, a short-term TSC evaluation method for low-voltage flexible interconnected distribution networks based on temporal convolutional network (TCN) is proposed, which combines speed and accuracy. Firstly, short-term TSC analysis of low-voltage flexible interconnected distribution networks is conducted and establish a data-driven short-term TSC evaluation architecture is established. Then, a model-driven short-term TSC evaluation method is adopted to obtain short-term TSC values to obtain training samples. The TCN model is then trained offline to obtain a nonlinear mapping between short-term TSC values and influencing factors with certain temporal characteristics. This enables the online evaluation of short-term TSC by the TCN system. Finally, the effectiveness of the proposed method is verified through system testing of low-voltage flexible interconnected distribution networks.