To provide information for precision management in tomato production under protected environments, the relationship between environmental factors and fruit growth of facility-cultivated tomatoes were studied and the optimum range of key environmental variables promoting fruit growth was identified. The temperature, light intensity, relative humidity and soil temperature at depth of 10cm in tomato cultivation facilities were recorded at one hour interval. The variation of daily tomato fruit growth were modeled from the data of fruit diameter increase by measuring transverse and longitudinal diameters of fruit every seven days. Data on seven environmental variables and fruit growth were analyzed, and a regression model was established with the DPS software package based on stepwise regression of key environmental variables and tomato fruit growth. It was shown that, the dynamics of environmental factors over time in two plastic sheds for spring tomato crops was similar. In both cases, there was a gradual increase of temperature, decrease of day/night temperature difference, and fluctuation of light intensity and relative humidity. In contrast, in two greenhouses with autumn tomato crops, there was a gradual decrease of air temperature but insignificant day/night temperature difference. Daily fluctuation of light intensity and relative humidity was similar to that in the plastic sheds in the spring. There was a great dependency among the seven environmental variables. For instance, the soil temperature rose with elevated air temperature while the relative humidity dropped with increase of light intensity. Daily fruit growth in the cultivars Fen Guan and Jin Peng increased dramatically first and decreased afterwards, which, however, fluctuated greatly in the cultivar Zhen Qi. The key environmental factors affecting daily fruit growth were significantly different among the three cultivation facilities.A regression model between daily tomato fruit growth and key environmental variables was developed for each of the three cultivars, Fen Guan, Jin Peng and Zhen Qi. The goodness of fit, as evidenced by the correlation coefficients, which were 0.9866, 0.9107 and 0.9237, respectively; and the F-statistics values, which were 29.32, 9.73 and 7.26, respectively for the regression function indicated that this model could be used for analysis and predication of tomato fruit growth with reasonable accuracy and precision. The optimum range of the seven environmental variables was defined with regard to their effects on promoting tomato fruit growth.