Nowadays, various methods are used to automatic count and recognize the blood cells in microscopic images. In detection and diagnosis of illnesses, the number of cells in the blood and determination of types of these cells is very important topics. 5 types of white blood cells, a red blood cell and platelet cell are located in human blood. In this paper, a new algorithm based on directed vector method is presented for blood cells recognition in images. In the pre-processing step of this algorithm, some image processing operators are used such as noise removal, thresholding and edge extraction methods. In feature extraction step, 8-way vector chain code and statistical characteristics are used. In classifier step, multilayer perception (MLP) artificial neural network (ANN) is used. The obtained results are satisfactory