Carbon stars are critical for understanding stellar evolution and exploring cosmic chemistry. This paper provides 7809 carbon star candidates identified from the Large-Area Multi-Object Fiber Optic Spectroscopic Telescope (LAMOST) DR11 and Sloan Digital Sky Survey (SDSS) DR18, using a parallel carbon star feature recognition method. This method deploys a parallel multi-interval feature representation model and a k-nearest neighbor classification model, effectively characterizing local and global molecular bands of carbon stars and enhancing feature discrimination, especially for weak features that are highly susceptible to noise interference. The recognition performance F1-score can exceed 96%. Moreover, with parallel strategies for interval representation, the search efficiency has improved. And we construct a new catalog of carbon stars subtypes, including 1616 barium (Ba) stars, 1720 C-H stars, 1621 C-N stars, 1205 C-R stars, and 1647 stars labeled as "Unknown" due to their low signal-to-noise ratios (S/Ns). Compared with the LAMOST DR11 pipeline, carbon stars by Li+ 2024, J/ApJS/271/12 and the SDSS DR18 pipeline, we identify 1403 and 1644 new candidates from LAMOST, and 661 new candidates from SDSS DR18, respectively. In our catalog, the majority of C-N, C-R, and Ba stars from LAMOST are located at low Galactic latitudes with |b|=30{deg}, while most candidates from SDSS are located at high Galactic latitudes. And the vast majority of Ba, C-H, C-R, and Unknown stars show relatively high effective temperatures. Some carbon stars exhibit stellar activity, potentially influenced by dense interstellar materials, internal magnetic fields, and rotation effects. In the g, r, and i bands, C-N and Unknown stars from LAMOST exhibit lower S/Ns, influenced by high interstellar density at low Galactic latitudes, while Unknown stars from SDSS show higher S/Ns. Among the newly discovered carbon stars, Ba from LAMOST, Unknown from SDSS, and C-R stars have higher S/Ns than other subclasses.