The quick development of mass customization in the apparel industry leads to an exponential increase of garment size combinations for markers, which induces a heavy and complex workload of marker making. In this context, due to the complexity of the problem, the classical marker making methods using the existing commercialized software are less performant in terms of efficiency and accuracy. Therefore, machine learning techniques, usually taken as efficient tools for extracting relevant information from data measured in uncertain and complex scenarios, are considered much simpler and faster. In this study, we apply the methods of multiple linear regression (MLR) and radial basis function neural network (RBF NN) to estimate marker lengths that are used in various garment production modes by considering various sets of garment sizes and different marker types. The experimental results show that the proposed approach leads to a good performance in estimating marker lengths of different types of markers (mixed marker and group marker) with diverse size combinations taken from various sets of garment sizes in both mass production and mass customization conditions.