Predicting the thermal transients of district heating (DH) network is the key to simulation analysis and operation optimization of DH system. Numerical methods can provide accurate prediction and sufficient information of thermal transients. But the high computation burden restricts the application of numerical methods, especially when applied to operation optimization of large DH networks. This dilemma can be relieved by suitably increasing the scales of time and spatial steps, but do not obviously affect the precision of the numerical models. However, there are few researches concerning such topics. In this paper, the optimal scales of time and spatial steps of a newly proposed implicit upwind model and the characteristic line model were studied for fast and accurate calculation. Results show that both models can ensure the prediction errors of the pipeline outlet temperature within +/- 0.5 degrees C. For characteristic line model, the recommended time step and spatial step are 60 s and 170 m < Delta x < 520 m. And for implicit upwind method, the recommended time step and spatial step are 20 s and 30 m. Besides, the implicit upwind model is unconditionally stable and provides more information on temperature distribution along the pipeline, especially when small and fast propagation occurs.