This study was aimed at searching new instruments for primary differentiation of dyspnea associated with lung or heart pathology. The study consisted of 3 stages and included analysis of "glossary dyspnea" of 482 patients aged 18-90yr. Stages i and II involved 359 patients with the known diagnosis: chronic obstructive pulmonary disease (COPD) (n=114, 31.8%), bronchial asthma (BA) (n=134, 3 7.3%), chronic heart failure (CHF) unrelated to lung pathology (n=111, 30.9%). The patients were 205 (57.1%) men and 154 (42.9%) women. Stage III involved 123 patients complaining ofdyspneapresumably ofpulmnonary or cardiac origin. 75 (61.0%) of these patients were men, 48 (39.0%) women aged 18-89 yr. Subsequent comprehensive diagnostics revealed BA in 47 (38.2%), COPD in 46 (37.4%), CHF in 30 (24.4%) patents. Stage I was designed to evaluate the dyspnea language based on the frequency of the phases of glossary dyspnea (Simon et al., 1999) and Fisher's Z criterion. It permitted to create the modified glossary dyspnea (MGD) consisting of 11i phrases and 7 clusters that revealed statistically significant differences between patients with BA, COPD, and CHE At stage II, we developed the model for estimating the probability of a priori diagnosis (PPD) as a possible cause of dyspnea based on the data of MGD and modified Bayesian classificator. At stage III, the PPD model was verified using blindly selected patients. Prior to verification, the patients filled MGD. The final diagnosis was specified in accordance with the national recommendations. on diagnostics of BA, COPD and CHE Comparison of the frequency of coincidence between preliminary (a priori) and definitive (a posteriori) diagnoses made it possible to evaluate sensitivity and specificity of the PPD model for BA, COPD, and CHF unrelated to lung pathology.