The recognition of Mathematical Expressions (ME) constitutes a challenging problem in character recognition research. A very few studies of offline Mathematical expressions have been so far reported in the literature. This paper focuses on offline handwritten and printed mathematical logical expressions recognition using Support Vector Machine classifier (SVM). In the work of expression recognition, the expressions were segmented into individual characters. The feature extraction method with combination of Normalized chain code and zone based density was used to get the features of a character. The present work considers logical expressions with subscripts for recognition. The experimental results for recognition rates of handwritten and printed expressions are reported. The result shows that the recognition rate of handwritten expression is 84.1% and that for printed expression is 90.3%.