In cities with serious air pollution, travel time and health damage significantly affect route choice by travelers (e.g., motorcycle and scooter drivers). Consequently, the classical Braess paradox is no longer realistic because it only considers the traveler’s value of time (VOT). In the current study, we describe a new transportation network paradox that considers both the VOT and the traveler’s perception of pollution damage. To examine the conditions that create the new paradox, we developed a novel method to compute a total comprehensive cost that combines the VOT with health damage. We analyzed the conditions for the new paradox and the system’s performance using a user equilibrium model and system optimization. Furthermore, an improved model is used to analyze how different transport modes influence the Braess paradox. We found that whether the new paradox occurs and the potential improvement of the system’s performance depend on whether the total travel demand falls within critical ranges. The bounds of these ranges depend on the values of the parameters in the function that describes the health damage and the link travel time function. In addition, high health damage significantly affects route choices and traffic flow distribution. This paper presents a new perspective for decision-making by transportation planners and for route choices in cities with serious air pollution.