Speech Recognition for desktop application is a system that allows the computer to identify and understand the commands spoken by a particular speaker based on individual information included in speech waves. The ultimate goal of work is to be able to produce a system that can recognize with expected accuracy, all commands that are spoken by particular person so that each operation of the application (power point) is accessed by corresponding commands. In the proposed work, Linear Predictive Coding (LPC) and Artificial Neural Network (ANN) are the techniques used for speech recognition system. Speech signals are taken and sampled directly from microphone and then they are processed using LPC method for extracting the features of speech signal. These features are used to construct the database to train the Artificial Neural Network using Back propagation method. Further the testing is conducted for the particular person by giving the voice command and if command is match/found then corresponding operation of application(power point) get enabled. The overall recognition accuracy of the proposed system with static or standard database is about 91.75% when tested with minimum 100 voice sample commands of a particular speaker. But, for real-time system, the accuracy is depending on the environmental situation and it will gives around 85-86% accuracy in worth.