Abstract
In the introduction of Javanese and Balinese script patterns, there are some research that have been conducted to determine the right method with different levels of accuracy. Some factors are very influential in determining the final result of research, such as the method used, research data, etc. This research uses two methods, namely the integral projection for feature extraction and neural network for classification. The purpose both of methods is to determine the accuracy of the result in recognizing the handwritten of the Sasak script This research is carried out to determine the effect of the number of nodes in a hidden layer. The highest accuracy if found in the use of hidden layers, including 21 nodes for the first hidden layer, and 14 nodes for the second hidden layer. The level of accuracy of this research is 41,38%.