Adequate greenhouse environmental management is very important for improving resource use efficiency and increasing vegetable yield. The objective of this study was to explore suitable climate and cultivation management for cucumber to achieve high yield and build optimal yield models in semi-closed greenhouses. A fruit cucumber cultivar Deltastar was grown over 4 years in greenhouse and weekly data of yields (mean, highest and lowest) and environmental variables, including total radiation, air temperature, relative humidity, and carbon dioxide (CO2) concentration were collected. Regression analyses were applied to develop the relationships and build best regression models of yields with environmental variables using the first 2 years of data. Data collected in years 3 and 4 were used for model validation. Results showed that total radiation, nutrient, temperature, CO2 concentration, and average nighttime relative humidity had significant correlations with cucumber yields. The best regression models fit the mean, lowest, and highest yields very well with R2 values of 0.67, 0.66, and 0.64, respectively. Total radiation and air temperature had the most significant contributions to the variations of the yields. Our results of this study provide useful information for improving greenhouse climate management and yield forecast in semi-closed greenhouses.