Crop yield changes regularly along with technological progress and climate change. Technological progress usually leads to the increase of crop yield, while climate fluctuations especially meteorological disasters often result in crop loss. In this paper, the long time-sequence crop yield was decomposed into the trend yield and the meteorological yield, which were respectively considered as the results of agricultural technology development and climate fluctuation. The trend yield could indicate the speed of technological progress and the potentiality of the crop output in a specific time interval. Meanwhile the meteorological yield could manifest the yield fluctuation that was resulted from the meteorological disasters, which could then be used to assess the potential risk of crop loss in a specific region. The results of yield decomposition of wheat, corn, soybean and paddy rice in main production countries of the world suggested that the yields increased rapidly in last 50 years for soybean and corn in the United States of America, Brazil and Argentina, for wheat in France, Germany and China, and for rice in China and Vietnam. Wheat yields of some European countries such as France and Germany, and rice yields of some Asian countries such as China and Thailand, had reached a peak and then decreased in recent years. As for as the meteorological risk assessment, besides crop yield reduction rate and its annual variation coefficient that were commonly used in studies, a new indicator of risk probability coefficient was built which was calculated based on the risk probability distribution function of crop loss using a weighted method. The risk probability distribution function was firstly obtained by a statistical method based on the relative meteorological yield series of a country. Standard normal transformation was then carried out, and the risk probability distribution could be divided into different intervals. The probability of each interval was assigned with different weight that increased progressively from low reduction rate to high reduction rate. The obtained integrated risk index(Pw) could indicate the risk degree of different crops in different regions. Three risk degrees i.e. low, moderate and high risk degree were divided, which were P_w≤2.0, 2.04.0, respectively. The results suggested that the meteorological risks of wheat in Canada and Australia, corn in America, and soybean in Brazil and Argentina were high and the P_w was more than 4.0. The risks of wheat in Russia and China, corn in Brazil and Argentina, soybean in America and China, and rice in India were moderate. The P_w values of wheat in America, Germany, France and India, corn in China, and rice in China, Vietnam and Thailand were less than 2.0, which meant their meteorological risks were low. The method in this paper can be helpful to evaluate the change trend of crop yield and assess the meteorological risk of agricultural production at global scale.