This paper presents a new epidemic model, SIDARTHE, which distinguishes between diagnosed and undiagnosed infected individuals and between different severity levels of illness. The model is designed to predict the trend of the COVID-19 epidemic in Italy, where the virus has become a significant health crisis. The SIDARTHE model is an extension of the classical SIR model, incorporating additional compartments to account for diagnosed and undiagnosed cases, as well as different severity levels of illness. The authors estimate the model parameters using data from February 20 to March 12, 2020, and simulate various countermeasure scenarios to assess their impact on the epidemic's progression. The results show that strong social distancing measures are necessary to control the spread of the virus, with mild measures leading to a delayed but less severe peak in infections and ICU admissions. The model also highlights the importance of distinguishing between diagnosed and undiagnosed cases to better understand the case fatality rate and the true extent of the epidemic. The SIDARTHE model provides a valuable tool for predicting and managing the COVID-19 outbreak, emphasizing the need for effective control strategies to mitigate the impact of the virus.This paper presents a new epidemic model, SIDARTHE, which distinguishes between diagnosed and undiagnosed infected individuals and between different severity levels of illness. The model is designed to predict the trend of the COVID-19 epidemic in Italy, where the virus has become a significant health crisis. The SIDARTHE model is an extension of the classical SIR model, incorporating additional compartments to account for diagnosed and undiagnosed cases, as well as different severity levels of illness. The authors estimate the model parameters using data from February 20 to March 12, 2020, and simulate various countermeasure scenarios to assess their impact on the epidemic's progression. The results show that strong social distancing measures are necessary to control the spread of the virus, with mild measures leading to a delayed but less severe peak in infections and ICU admissions. The model also highlights the importance of distinguishing between diagnosed and undiagnosed cases to better understand the case fatality rate and the true extent of the epidemic. The SIDARTHE model provides a valuable tool for predicting and managing the COVID-19 outbreak, emphasizing the need for effective control strategies to mitigate the impact of the virus.