An intelligent system for non-isothermal drying of medicinal plants, based on machine vision, sensor fusion and neural network was developed. Air temperature, velocity and humidity, along with material size and moisture content were inputs to the neural model for diffusivity. Temperature, time, mass, volume and colour were inputs to the neural model for quality. Isothermal low-temperature drying of ginseng root and blueberry showed extremely low effective diffusivity (0.2 -0.75)*10 -10 m 2 /s. In contrast, nonisothermal drying demonstrated a potential to increase diffusivity and prevent quality losses. Testing of the intelligent drying system showed reduced drying time from 240 to 60 hours for ginseng, and from 110 to 30 hours for blueberry with desired product quality.