We examine the feasibility of detecting auroral emission from the potentially habitable exoplanet Proxima Centauri b. Detection of aurorae would yield an independent confirmation of the planet's existence, constrain the presence and composition of its atmosphere, and determine the planet's eccentricity and inclination, thereby breaking the mass-inclination degeneracy. If Proxima Centauri b is a terrestrial world with an Earth-like atmosphere and magnetic field, we estimate the power at the 5577Å OI auroral line is on the order of 0.1 TW under steady-state stellar wind, or ∼100× stronger than that on Earth. This corresponds to a planet-star contrast ratio of 10 −6 − 10 −7 in a narrow band about the 5577Å line, although higher contrast (10 −4 − 10 −5 ) may be possible during periods of strong magnetospheric disturbance (auroral power 1 − 10 TW). We searched the Proxima Centauri b HARPS data for the 5577Å line and for other prominent oxygen and nitrogen lines, but find no signal, indicating that the OI auroral line contrast must be lower than 2 × 10 −2 (with power 3,000 TW), consistent with our predictions. We find that observations of 0.1 TW auroral emission lines are likely infeasible with current and planned telescopes. However, future observations with a space-based coronagraphic telescope or a ground-based extremely large telescope (ELT) with a coronagraph could push sensitivity down to terawatt oxygen aurorae (contrast 7 × 10 −6 ) with exposure times of ∼1 day. If a coronagraph design contrast of 10 −7 can be achieved with negligible instrumental noise, a future concept ELT could observe steady-state auroral emission in a few nights.