Abstract: Based on the relevant data of Shenyang real estate market from 2003 to 2009， the data from 2010 to 2011 were forecasted using the generalized regression neural network with a smoothing factor of 01 which has excellent approximation， and were compared with the true data. The results show that the developing investment， housing average price， and vacant areas of the real estate in Shenyang are in high level. Moreover， the real estate market is still in the boom， but belongs to the later period. Government， developer， and home buyer should pay attention to the real estate bubbles which may emerge in the future.

Key words: real estate period, radial basis function, generalized regression neural network, real estate bubble, sustainable development