Due to the obvious inter-annual variation of wind speed

the long-term resource status of the site area needs to be evaluated in the process of feasibility analysis of wind farm project construction. Based on measurement-correlation-prediction（MCP）

a meta-learning fusion evaluation method considering multi-area meteorological characteristics is proposed in this paper. The observation information of wind speed

wind direction

air temperature and air density of several reference meteorological stations is combined with the same period wind speed information of the wind farm to be evaluated

so that the long-term wind condition information of the wind field to be evaluated is obtained

and then the long-term wind energy reserves of the wind farm to be evaluated are calculated. Firstly

the Pearson correlation coefficient is used to analyze the wind speed correlation between the target wind farm and the reference station. Then

the feature factor method is used to select key characteristics to avoid redundant information from affecting the accuracy of the results and the efficiency of calculations. Eventually

different ensemble models

i.e.

the Random Forest model and the XGBoost model

are used to establish short-term correlate model. Comparison of predicted results and actual observation is used to verify the validity of proposed approach. The research results show that the introduction of multi-dimensional meteorological features can effectively improve the fitting effect of the long-term wind speed of the wind farm to be built

and then improve the accuracy of the evaluation results of wind energy resources to meet the needs of practical projects.