In this paper, we present an effective quality assessment method based on the relation intensity ratio and detail similarity for image quality assessment (IQA) with the full reference image, which first allows us to compute the nonlinear gradient magnitude with Gaussian smoothing of the reference and distorted images and construct the relation intensity ratio and detail similarity between them. Next, the final IQA map is formed by linearly combining the relation intensity ratio with the detail similarity. Finally, we adopt a new pooling strategy which effectively integrates the mean and standard deviation of the final IQA map to accurately predict image quality. Experiments based on two publicly available databases show that the proposed method can provide accurate predictions compared with most state-of-the-art IQA methods.