Evaluating the balance between a city’s physical and socioeconomic environmental development is crucial for creating sustainable and livable urban spaces. Although they might appear contradictory, they jointly support the comprehensive sustainable urban development strategy. Traditional methods usually focus on assessing this balance from a specific perspective, such as how neighborhood greenery shapes real estate value. Yet, they fail to deliver a holistic balance assessment in developing the physical and socioeconomic dimensions. To fill this gap, this study introduces a research framework that measures this balance through house prices based on transferred bias. Using house price as an indicator shaped by both physical and socioeconomic environments, the framework first constructs a series of deep learning models to estimate house prices through street view images for each city. These models capture the relationship between neighborhood appearance and house price. Second, by leveraging transfer inference, we introduce neighborhood appearance from one city into the model trained from another city. This process identifies the transferred bias, which is the disparity between inaccurate inference resulting from a mismatched neighborhood appearance and the trained model. Through transferred bias, we can quantify the differences in physical and socioeconomic environments across cities and evaluate the urban balances of these two environments. The results show that the transferred bias effectively quantifies the disparities among cities in physical and socioeconomic environments, thereby facilitating further investigation into the urban balance between these two environments.