Few questionnaires are available for routine assessment of dyspnea. The study aimed to design a self-administered questionnaire assessing the impact of chronic dyspnea on daily activities, named DYSLIM (Dyspnea-induced Limitation). The development followed 4 steps: 1: selection of relevant activities and related questions (focus groups); 2: clinical study: internal and concurrent validity vs. modified Medical Research Council (mMRC), Baseline Dyspnea Index (BDI) and Saint George Respiratory Questionnaire (SGRQ); 3: item reduction; 4: responsiveness. Eighteen activities (from eating to climbing stairs) were considered with 5 modalities for each: doing the task slowly, taking breaks, seeking assistance, changing habits, and activity avoidance. Each modality was graded from 5 (never) to 1 (very often). Validation study included 194 patients: COPD (FEV1 ≥ 50% pred: n = 40; FEV1 < 50% pred: n = 65); cystic fibrosis (n = 30), interstitial lung disease (n = 30), pulmonary hypertension (n = 29). Responsiveness was evaluated by post-pulmonary rehabilitation data in 52 COPD patients. Acceptability was high and short term (7 days) reproducibility was satisfactory (Kappa mostly above 0.7). Concurrent validity was high vs. mMRC (Spearman correlation coefficient, r = 0.71), BDI (r = - 0.75) and SGRQ (r = - 0.79). The reduced questionnaire with 8 activities (from cleaning to climbing stairs) and 3 modalities (slowly, seeking help, changing habits) showed a comparable validity and was chosen as the final short version. Effect size of rehabilitation was good for both the full (0.57) and short (0.51) versions. A significant correlation was also found between changes of SGRQ and DYSLIM post rehabilitation: r = - 0.68 and r = - 0.60 for full and reduced questionnaires, respectively. The DYSLIM questionnaire appears promising for the evaluation of dyspnea-induced limitations in chronic respiratory diseases and seems suitable for use in various contexts. •Dyslim is a self-administered questionnaire assessing dyspnea-induced limitation.•Development of the questionnaire followed modern psychometric rules.•Dyslim is unidimensional, reproducible, responsive, with a high concurrent validity.•The short version comprises 8 questions with 3 modalities, requiring less than 3 min.