Abstract: Agricultural insurance is an important policy tool to protect farmers’ income and to stabilize agricultural production. Based on the yield and price data of tomato, cucumber, bell pepper, and cabbage in large and medium-sized cities from 2004 to 2016, and applying the semi-parametric Copula method, this paper analyzed the edge distribution of vegetable yield fluctuation, price fluctuation, and the joint distribution function of yield and price fluctuation. In addition, this paper also examined the correlation degrees between output and price fluctuations and the pure premiums of production insurance, price insurance and income insurance under different coverage levels. Results show that: 1) the t-Copula function is more suitable for fitting the joint distribution of vegetable yield fluctuations and price fluctuations than other Copula functions; 2) there is a negative correlation between the yield and price of vegetables, and the productivity risk and price risk of vegetables can be partially hedged; 3) different vegetables are suitable for different types of agricultural insurance programs. Price insurance is suitable for varieties with weak correlation between yield and price, while income insurance is suitable for varieties with strong negative correlation between them; and 4) at a level of 70 to 90 percent, income insurance premiums of tomato, cucumber, bell pepper and cabbage in large and medium-sized cities are 7.97%~21.08%, 4.03%~14.41%, 3.90%~13.04% and 1.45%~9.32%, respectively.