How to achieve efficient appreciation and intelligent parsing of Tang and Song literatures are two major difficulties in the field of studying Tang and Song literatures nowadays. Based on this, this paper investigates the application of multidimensional spatio-temporal IoT parsing model combined with cone stacking algorithm in the appreciation of Tang and Song literatures. Firstly, we summarize the current research status of Tang and Song literature appreciation, and propose a multidimensional spatio-temporal IoT analysis model based on cone stacking algorithm according to the characteristics of Tang and Song literature, and combine the application ideas of various appreciation models in Tang and Song literature appreciation to normalize the multidimensional analysis of traditional Tang and Song literature database to achieve accurate appreciation from different perspectives, and then analyze the accuracy and objectivity of the appreciation model. The accuracy and objectivity of the appreciation model are then analyzed. Secondly, the set of evaluation functions of the multidimensional spatio-temporal IoT analysis model in this application is constructed, and the characteristics and appreciation ideas of current literary masters are combined in the optimal appreciation, and the error of the composite appreciation analysis function is improved, while the cone stacking algorithm is combined with the multidimensional spatio-temporal IoT analysis appreciation model to realize the internal intelligent learning. Finally, a validation experiment is designed, and the results show that the multidimensional spatio-temporal IoT analysis model based on the cone stacking algorithm can quickly and accurately achieve the appreciation of conventional Tang and Song literature from multiple levels.