It is a well-known fact that building character recognition systems is one of the hottest areas of research as it is shown over the Internet. Due to its wide range of prospects, Character recognition has been an important part of the civilization for many centuries. We depend on digitally assembled and recognizable characters for common communications. This paper describes a method to recognize on-line handwritten Kannada characters, the language used by the state of Karnataka. A subset of the Kannada alphabet was chosen for the study. The Back propagation neural network is used for recognition.