The facility light environment, including facility light intensity and light quality, is an important factor affecting the growth and development of crops. There is a significant interaction between the light intensity, light quality and photosynthetic rate at different temperatures. It is one of the most urgent problems for facility agriculture to establish an intelligent regulation model of light environment for facility cucumbers with light quality demand, and effectively improve the light environment of crops. A multifactor nesting experiment was designed to obtain multidimensional sample data, and a support vector regression algorithm photosynthetic rate prediction model was constructed, which coupled temperature, light intensity, and light quality. Then, based on the particle swarm optimization algorithm, the optimal light intensities and light qualities under specific temperature conditions were obtained quickly. Finally, based on the optimization results, the intelligent regulation models of red and blue light were constructed by partial least squares regression method. As a result, the fitting degrees of training set and test set of the photosynthetic rate prediction model were 0.9971 and 0.9969, respectively, and the root mean square errors of training set and test set were 0.3630μmol/(m2·s) and 0.4367μmol/(m2·s).The root mean square errors of the intelligent regulation models of red and blue light were 15.0878μmol/(m2·s) and 10.1383μmol/(m2·s), respectively. Compared with the traditional fixed light quality models, the regulation effect of the model was significantly improved, which indicated that these models provided an important basis for the effective regulation of the light environment of facilities.