In this paper, we have explained the different approaches of text-to-speech conversion for Kannada language using Neural Networks. Text-to-speech conversion (TTS) has a wide variety of applications like it serves as an effective aid to the visually impaired by offering a computer-generated spoken voice that would “read” text to the user. TTS can also be used as an interactive educational appliance, as an assistant, etc. The application field of synthetic speech is expanding fast while the quality of the TTS system is also increasing steadily. Text-To-Speech conversion using Neural Networks is found to be more accurate when compared to many other methods like Concatenation, Digital Signal Processing, Articulatory Synthesis, etc. The different approaches of neural networks that we have mentioned in this paper explain its own methodologies, its applications and also its pros and cons.