To resolve problems of information asymmetry and low matching efficiency in freight market, a freight resource sharing platform must provide accurate and readily acceptable vehicle-cargo matching results. We study a two-sided matching model, and analyze the impact of suppliers and demanders’ loss aversion on matching results. We find that users’ loss aversion positively influences the accuracy of vehicle-cargo matching. Next, we construct an evolutionary game model and investigate the trend of users’ strategic selection of the matching results recommended by the platform and the factors influencing their selection. Within their respective limits, the platform’s service level, the users’ initial acceptance probability and waiting cost are found to have a positive impact on the evolutionary trend of users’ acceptance of the matching results. The models are verified using numeric analyses. Several suggestions are made for improving platform matching efficiency.