The technology, based on the neural networks of the human brain, maintains intonation and cadence to mimic the original speaker's voice.
Microsoft teamed up with researchers from the University of Toronto to develop the tech, which reached a breakthrough two years ago. The computing giant claims it offers 30% more accurate translation than previous methods.
"We hope in a few years that we'll be able to break down the language barrier between people," said Microsoft Research's Rick Rashid.
At present, one word in seven or eight is translated incorrectly, compared to the ratio of one in five of existing techniques.
"Of course, there are still likely to be errors in both the English text and the translation into Chinese, and the results can sometimes be humorous," Rashid added. "Still, the technology has developed to be quite useful."
Microsoft is yet to announce whether the software will be made commercially available.