We present the problem of risk estimation/assessment of the new coronavirus (COVID-19) spread in Poland. In the literature, a method is given to explain the COVID-19 disease and try to predict the occurrence of the next wave of COVID-19. A Gaussian composite model is used in literature to identify COVID-19 disease. In this model, however, the prediction of the upcoming waves is dependent primarily on the set of selected samples. For different sets of samples, distinct timings of the upcoming waves are obtained. In this article, we assume that probability is associated with each set of samples, which are used for the prediction of the upcoming waves. By exploring this aspect, it is shown that a better prediction of the upcoming COVID-19 waves can be performed.