Since SVM is very sensitive to outliers and noises in the training set, a fuzzy support vector machine algorithm based on affinity among samples is proposed in this paper. The fuzzy membership is defined by not only the relation between a sample and its cluster center, but also those among samples, which is described by the affinity among samples. A method defining the affinity among samples is considered using a sphere with minimum volume while containing the maximum of the samples. Then, the fuzzy membership is defined according to the position of samples in sphere space. Compared with the fuzzy support vector machine algorithm based on the relation between a sample and its cluster center, this method effectively distinguishes between the valid samples and the outliers or noises. Experimental results show that the fuzzy support vector machine based on the affinity among samples is more robust than the traditional support vector machine, and the fuzzy support vector machines based on the distance of a sample and its cluster center.