Accurately determining the daily theoretical line loss rate in low-voltage distribution networks is challenging, making it difficult to quantitatively assess its fluctuation range. To address this issue, this paper proposes a probabilistic analysis method for evaluating the daily theoretical line loss rate in low-voltage distribution networks. First, the actual three-phase four-wire connection of the low-voltage distribution network is considered, and it is assumed that the random models of three-phase voltage, distributed power output, and the three-phase active and reactive power of the distribution transformer on the low-voltage side are known and simulated. Then, source-load correlation is incorporated, and Latin hypercube sampling, combined with the equal probability transform principle and rank correlation, is employed. Finally, using the Monte Carlo simulation method, the probabilistic distribution of power flow and the daily theoretical line loss rate is calculated by the Newton injection current method, taking three-phase unbalance into account. This approach provides a decision-making basis for reducing losses in low voltage distribution networks. Experimental data from the Hengshan Garden low-voltage distribution network validate the effectiveness of the proposed method.