An extensive experimental database consisting of 2838 shrinkage data points and 3598 creep data points is used to evaluate the accuracy of the newly proposed CEB-FIP 2010 model in predicting the creep and shrinkage of concrete structures. To study the applicability of the model for high-strength concrete in general environments, the database was developed by only retaining the test data of concrete components with the average compressive strength greater than 30 MPa and the relative humidity in the test environment less than 95%. On this basis, combined with the B3 and CEB variation coefficient methods, the paper mainly adopts the residual method to assess the accuracy of the CEB-FIP 2010 model and compare it with the previous model, CEB-FIP 1990. The influences of several properties, such as the compressive strength, the age of concrete, the relative humidity, and the component size on the prediction accuracy of these two models are further studied. The results show that for the CEB-FIP 2010 model within the time interval of 0–9000 days, 52% and 48% of the shrinkage strain residuals of the total specimens are located in the negative and positive regions, respectively, while the positive and negative regions of the CEB-FIP 1990 model account for 73% and 27%, demonstrating the CEB-FIP 2010 model has better performance in predicting shrinkage strain than the CEB-FIP 1990 model, whereas the two models have comparable accuracy in predicting creep compliance. The CEB-FIP 2010 model is more reliable for considering the effects of compressive strength, relative humidity, and age at loading on shrinkage and creep than for considering the effect of member size.