Multimedia services are typically much more sensitive to throughput, delay, and packet loss than traditional services. These parameters give technical performances of the network but they do not reveal the quality of service (QoS) perceived by users. Hence, how to capture users’ overall perceptions when they are using network services is an important issue. This paper designs and implements a QoE measurement framework to capture users’ perceptions. The concept of framework is first to create a sample database in the analysis center, and then use the machine learning algorithm to train the sample database. Finally, a QoE satisfaction model is built to estimate the user QoE satisfaction. Users in any enterprise can use the designed QoE measurement interface to measure their QoE. When the users have a low QoE, a service provider can use the built QoE satisfaction model to find quickly the reasons that degrade QoE. The implementation results showed that the number of clicks on grading buttons and throughput exhibit similar behavior, the created sample database has a higher reliability, and the build QoE satisfaction model produces high accuracy rates.