Abstract: A cabinet with filter life prediction function and a method of predicting filter life are provided. The cabinet has an airflow module generating airflow flowing through a filter module. An air quality sensor continuously monitors the air quality sensing value of the airflow and records the sensed value in association with the sensing time. A regression analysis is performed based on the record data to obtain a regression model when a modeling condition is met. When a prediction condition is met, a time period for the currently sensed value to decrease to a threshold is calculated to be a predicted remaining life of the filter module. A notification is generated when the predicted remaining life of the filter module is less than a life threshold.