The prediction method of MUF (Maximum Usable Frequency) has been extensively deliberated as a pivotal technology in HF (High Frequency) communication. To verify the reliability of MUF prediction models commonly used in HF communications, we evaluated the prediction accuracy of seven typical prediction models based on maximum observed frequency (MOFs) and predicted MUFs. The oblique-sounding (OS) circuit was from Changchun to Jingyang, with the vertical-sounding (VS) station situated in Beijing which is, approximately at the midpoint of this oblique-sounding (OS) circuit. The RMSE (root mean square error) and RRMSE (relative root mean square error) between the MOFs and predicted MUFs by the model were used to evaluate the accuracy of the models. The results of statistical analysis show that: (a) the prediction results of the seven models are consistent with the observed results in trend, among which the prediction results of the Lockwood model (LWM), the nonuniform ionospheric conditions MUF calculation model (NYM), and the model proposed by Istituto Nazionale di Geofisica e Vulcanologia (INGV) are in good agreement with the observed results; (b) In the four seasons, the seasonal distribution of the seven models is significantly different, and the INGV model has a better prediction effect; (c) The INGV model is superior to other models in relatively high solar activity epochs and relatively low solar activity epochs, followed by the LWM model, with little difference between the two models and close prediction accuracy; (d) The INGV model has a better prediction result in four different periods of a day, and the improved simplified MUF calculation model (MSPM) has a great prediction result in the periods of nighttime.