The two-terminal fault location theory based on the lumped parameter model is widely used in transmission line protection. For short and medium distance transmission lines,the model error is negligible. However,there are still other factors that affect the ranging results. It is easy to produce large error if the fault is removed in a very short time. In order to improve the accuracy and universality of ranging,the absolute error of the two-terminal fault location algorithm based on the lumped parameter model under the common influencing factors is analyzed. By using orthogonal experiment design and simulation calculation,the influence degree of different influencing factors on ranging error is compared. It is found that data window selection is the most influential factor. Besides,the ranging error of the proximal fault of the line,compared to other places,is greatly affected by the influencing factors. According to the two features,a practical algorithm of adaptive optimal data window based on segmented processing is proposed. By verification,the influence of ranging error calculated by the improved algorithm is greatly reduced.