This paper builds a spatial design model based on the reconstruction of 3D visual spatial points in an indoor environment and combines the minimum Euclidean distance to calculate the vector relationship between spatial points and realize the corresponding spatial point matching. For the exploration of traditional elements in spatial design, an extraction method based on the improved cuckoo algorithm is proposed. By adding a dynamic inertia weight adjustment method, one inertia weight value is randomly selected from the Gaussian distribution each time to balance the algorithm’s local and global search. By extracting traditional elements, calligraphy is the most frequent in the home space, with 79.4% integration. The tea set is the most frequent traditional element in the dining space, with 84.53% integration. A couplet is the most frequent traditional element in the public open space, with 71.98% integration. The extraction based on intelligent algorithms can help designers better integrate traditional elements in the design of environmental spaces.