With the rapid development of low-altitude economy, the number of aerial vehicles in urban low-altitude airspace has increased dramatically, and the risk of collisions between unconscious aerial objects such as free-floating balloons and unmanned aerial vehicles (UAVs) has become increasingly prominent. To address this issue, this paper constructs a comprehensive collision probability assessment model for UAVs and free-floating balloons, taking into account the characteristics of urban low-altitude wind fields, the dynamic and kinematic models of free-floating balloons, and factors such as von Kármán turbulence disturbances. Firstly, through a multi-model fitting comparison of measured wind field data, the combination of the wind speed power law model and the wind direction linear model was determined as the optimal expression for establishing the average wind field. Based on the principles of fluid mechanics, a dynamic equation for free-floating balloons was established, incorporating major forces such as buoyancy, added mass effect, and air resistance. The von Kármán energy spectrum and additional vertical gusts were used to describe the characteristics of low-altitude atmospheric turbulence. On this basis, a collision probability assessment framework based on the Monte Carlo method was proposed. Through large-scale random simulations, statistical indicators such as point estimates, standard errors, and confidence intervals for collision probability were provided. In the benchmark scenario, the point estimate of collision probability was 8.58%, with a 95% confidence interval of [8.03%, 9.13%] and a standard error of 0.28%, indicating that the model has high accuracy and reliability. Further parameter sensitivity analysis showed that increased turbulence intensity significantly reduces collision risk; UAV radius is strongly positively correlated with collision probability; while balloon payload is significantly negatively correlated with collision probability. In summary, the assessment framework established in this paper provides a quantifiable basis for autonomous collision avoidance and path planning of UAVs in environments containing free-floating balloons and other aerial objects, and also offers data support and methodological references for setting risk thresholds and safety control strategies for aerial objects in urban low-altitude airspace.