This article discusses the application of machine learning to the quantum regression of biogas production rate in anaerobic reactors. Anaerobic fermentation is a well-proven tool in wastewater treatment plants for the treatment of raw sludge. It can also be used for renewable energy sources by collecting biogas in anaerobic boilers. Installation operators usually set operating parameters, such as temperature, according to expert knowledge. In order to fully exploit the potential of operational management, in this study we calibrated a new temporal thermonuclear transformer based on data from the time series of the life scale over six years, along with categorical features such as public holidays. The design of the model ensures the interchangeability of the output data compared to traditional data methods that use more attention. In addition to predicting average biogas production rates over the next seven days, our model also provides quantiles, which makes it less susceptible to strong fluctuations. We used three well-known statistical methods as a guide. The average absolute percentage error of our predicted approach is less than 8%.