<p>Accuracy allocation is crucial in the accuracy design of machining tools. Current accuracy allocation methods primarily focus on positional deviation, with little consideration for tool direction deviation. To address this issue, we propose a geometric error cost sensitivity-based accuracy allocation method for five-axis machine tools. A geometric error model consisting of 41 error components is constructed based on homogeneous transformation matrices. Volumetric points with positional and tool direction deviations are randomly sampled to evaluate the accuracy of the machine tool. The sensitivity of each error component at these sampling points is analyzed using the Sobol method. To balance the needs of geometric precision and manufacturing cost, a geometric error cost sensitivity function is developed to estimate the required cost. By allocating error components affecting tool direction deviation first and the remaining components second, this allocation scheme ensures that both deviations meet the requirements. We also perform numerical simulation of a BC-type (<i>B</i>-axis and <i>C</i>-axis type) five-axis machine tool to validate the method. The results show that the new allocation scheme reduces the total geometric error cost by 27.8% compared to a uniform allocation scheme, and yields the same positional and tool direction machining accuracies.</p>