Abstract: Water supply is an essential to the normal growth of plants, due to too much or insufficient water can seriously confine the yield and quality of greenhouse crops. The traditional greenhouse irrigation method mainly triggers quantitative irrigation system through timing control and radiation thresholds, without considering the comprehensive impact of microclimate changes on water consumption of crops in the greenhouse environment. In the conventional model, such as converting crop transpiration to irrigation water demand with the empirical ratio factor, there is a lacking the fact that the growth of crop canopy leaves can weaken the effect of solar radiation on water consumption of plants below the canopy. The conventional water supply methods cannot meet the actual needs of crops, and thereby inevitably cause serious waste of water resources and environmental pollution by excessive discharge of nutrient solution waste liquid, eventually deteriorate the economic benefit of crop production in greenhouse. Therefore, it is of great practical value to establish an irrigation model based on crop transpiration physiological characteristics and intercepted radiation effects of actual canopy growth changes. The relevant research of the soilless substrate cultivation crop transpiration model pointed out that the surface evaporation of the substrate can be small or even negligible, indicating that the soil evaporation does not need to be considered when estimating the tomato transpiration of the soilless substrate cultivation. Penman-Monteith evapotranspiration model including soil evaporation cannot accurately estimate the transpiration of substrate-cultivated tomato in terms of greenhouse. In addition, some models were introduced a grass datum with a fixed crop height and a large area covering the ground, in order to empirically simplify the calculation of impedance parameters, and then used historical data to fit crop coefficient as a medium for converting reference evapotranspiration to actual evapotranspiration. However, the estimation process did not consider the actual growth of the crop evapotranspiration surface. In this study, a new transpiration model was established for the tomato plants cultivated on rice-husk charcoal substrate in the greenhouse. The calculation of empirical parameters was modified based on soil evaporation in the original representation of the evapotranspiration model, according to the distinction between soil-cultivated crop evapotranspiration and substrate-cultivated crop transpiration. TOMGRO model was introduced to simulate real-time growth changes of tomatoes to replace the grass datum of fixed growth parameters. The new transpiration model has added the resistance parameters of the water vapor exchange between the leaves and outside air, while maintaining the characteristics of crop transpiration, indicating more suitable for the greenhouse microclimate environment and the real-time growth of crops. In the simulated transpiration results of historical data from each full cycle of tomato growth in four years, compared with the greenhouse soil-cultivated crop evapotranspiration model, the transpiration model of substrate tomato shows a small error to simulate tomato transpiration, where the average absolute error are less than 0.1 mm/d, the average relative error are less than 10%, the root mean square error are less than 0.1 and the correlation coefficient are greater than 0.95. Considering that due to the net radiation interception weakening solar radiation, the transpiration model of substrate tomato can be used to describe the transpiration changes of crop canopy, but ignore the water consumption of the entire plant canopy and below. Therefore, a transpiration-radiation integrated irrigation model was established by multiple linear regression method, where introducing the solar radiation linear water supply model as a linear compensation to estimate the tomato irrigation water requirement using the substrate tomato transpiration model. Comparing the actual irrigation data, the average absolute error of the simulated irrigation volume are less than 0.29 mm/d, the average relative error are less than 20%, and the root mean square error are no more than 0.3. These errors are also less than the allowable error 0.3 mm/d of manual recording irrigation data in the experimental greenhouse. The estimation of greenhouse tomato water consumption in substrate cultivation scenario, transpiration-radiation integrated irrigation model have high accuracy, certain rationality and practicability. These models can provide a theoretical basis for the prediction of irrigation water consumption of rice-husk charcoal substrate cultivated tomato in the greenhouse and irrigation implementation plan.