Abstract: In order to improve the forecast accuracy of cloud landscape in mountainous scenic spot,the temporal change and the vertical distribution characteristics of some meteorological elements such as temperature,humidity and so on were first analyzed based on the observed cloud data from the Huangshan weather station and the NCEP/NCAR reanalysis data during 2004-2014.Then,the forecast factors reflecting the cloud landscape weather features were selected,and their thresholds and negative indexes were determined by analyzing the distribution characteristics of the forecast factors.Finally,the forecast model of cloud landscape at 08:00 during winter half-year was developed with index accumulation method and its total forecasting accuracy and T_S grade was 88% and 31%,respectively.In addition,the model was further verified using the data from 2015 to 2016.As a result,the forecast accuracy of cloud landscape is 88% and the total T_S grade is 36%.Specifically,T_S grades in March and October are 44% and that in January is 38%.In conclusion,the model simulation results can provide available reference to the forecast of cloud landscape in Huangshan Mountain in winter half-year.