Results of the UAV task allocation in the state of determinacy always become unfeasible or suboptimal when suffering from the uncertain factors in battlefield. The traditional expected value model is usually suitable for long-term planning, more difficult to consider the effect on certain decision making by the fluctuation of the uncertain variables. To deal with the uncertain programming in the UAV task allocation, the minimum-risk model is firstly established with a belief degree objective function based on uncertainty theory. Then, the model is transformed into an optimization problem with fractional objective by introducing two kinds of assumption. Finally, an auxiliary function is defined, and a ratio one-dimensional search exact algorithm is designed according to the monotonicity. The results of simulations show that the proposed method can keep the UAV away from the targets with larger standard deviation, and provide numerous alternative schemes with difficult belief degree levels by adjusting the target profit.