The existing methods for measuring Web service QoS (quality of service) are not accurate as they cannot precisely quantify the ambiguity of user preference while neglecting the characteristics of QoS data set. This paper presents a QoS measuring algorithm which employs subjective and objective weight. The new approach can automatically adapt to user preference with a subjective weight calculation method, and can evaluate service performance accurately with an objective weight calculation method. The algorithm takes a comprehensive consideration of both subjective and objective aspects to measure Web service QoS, therefore the measure results can conform to user preference and reflect the overall service performance accurately. The theoretical analysis and experimental results based on the QWS real data set show that the proposed algorithm can measure Web service QoS accurately.