Background: The novel virus (COVID-19) pandemic threats the global most since the World War II yet difficult to design effective policy to respond it because widely utilized models only support to predict the future tendency within a narrow time-window. Besides, existing pandemic metrics are sensitive to the detecting policy ( p <0.001), which make them hard to reveal the real infection phenomena, because the fluctuation of these metric values may be an outcome of a suddenly changed detecting policy instead of real infection increase or decrease. Methods: We developed a Benchmark, Amendment and Validation model (BAVM) to make the above complex phenomena clearer, and combined with the hypothesis of potential periodicity of COVID-19 to devote jointly the retreated prediction of pandemic within a broad time-window. Onward, we utilized 113 countries’ epidemic data from January 21, 2020 to August 31, 2020 and detecting policies to carry out our methodology, then validated the results through statistical tests involving chi -square test, one t-test and independent t test. Results: Results exhibit the 4-month of potential periodicity between two adjacent peaks of pandemic ( t =–1.56, p =12.4%). Besides, whether or not the peak emergence has no effect on COVID-19 dynamic trajectory, but the time to firstly peak affects. Meanwhile, uprising the quarantine rate exhibits the earlier expedition towards the first peak emergence (9.7% vs. 6.2%, p =4.1%). On the contrary, the delay of first peak increased the infection rate (0.6% vs. 0.3%, p =1.5%) but also the discharge rate (65% vs. 74%, p =3.8%). Moreover, the indication of the retreated tendency of COVID-19 pandemic is that the next peak should emerge but in fact vanished after one periodicity. Otherwise, the pandemic enters into the next worse phase, typical of high mortality (5% vs. 3.4%, p =5.3%) and low discharge rate (65.8% vs. 74.1%, p =4.2%). Conclusion: To recognize the retreated tendency of this pandemic is quite significant for either the global or the local policy-making, especially in the regular epidemic prevention, control measures and work resumption.