This study was performed to investigate the effectiveness of the aspiration trap method for collection of sputum by comparing with the conventional method which was collecting specimens at canular cap swab. In this study, the author tested by two methods to collect specimens from 46 patients who were cared with tracheostomy and intubation at the intensive care unit of an university hospital in Pusan, and investigated the incidence of the lower respiratory tract infection, the consistency between the two methods, the level of specimen contamination. Major results were as follows: Among the patients, 35 were cared with tracheostomy and 11 were cared with intubation. In clinical diagnosis we were classified the subjects in to two group, 17 of pneumonia group and 29 of non-pneumonia group. A total of 247 strains were isolated. Among them, most three strains were Serratia marcescens (62 strains; 25.1%), Pseudomonas aeruginosa (52 strains; 21.1%), and Acinetobacter baumannii (19 strains; 7.8%). Out of total, 188 (76.1%) strains were Gram negative bacilli. The isolated strains by the aspiration trap method were the average 2.1 strains, but by the canular cap swab method were 1.6 strains. In spite of the high contaminated possibility from the incision site and the oral cavity swab, the low isolated rates of the canular cap may be the dried environment of the canular of cap area. But the contamination rates were 57.2% of the canular cap, 51.5% of the oral swab and 50.5% of the incision site swab, respectively. The consistency of predominant microorganisms according to collection method were 86.7% of aspiration, 78.3% of canular, 74.3% of incision, and 63.6% of oral. In conclusion, the aspiration trap method fur the sputum collection from the patients with intubation of tracheostomy showed the lower contamination rate of the specimens and it was helpful for rapid, accurated interpretation of the lower respiratory tract infection and hospital infection.