Visual landscape quality represents a potential attribute of landscapes that affects people’s perception and psychological well-being. With the perceived sensory dimensions as a conceptual framework, this study proposes methods to measure visual quality in both real environments and virtual models, using image-based metrics from computer vision techniques and 3D model-based metrics from parametric modelling techniques. Using the Clementi Woods Park in Singapore as a case study, we compared these two types of metrics using statistical methods and proposed an approach of using a regression model from empirical studies to estimate subjective preference for design scenarios and thus to evaluate the result of landscape design scenarios.