Quantitative data on lung structure, such as volume, surface area and length, are used for assessment of the functional performance of the lung during normal development and inflammatory-related diseases such as chronic obstructive pulmonary disorder (COPD), and carcinogenesis, in animal models. Stereology is considered as the gold standard to obtain quantitative data on lung structure, with a key advantage being to quantify irregular three-dimensional structures on the basis of measurement made on two-dimensional sections. Therefore, preservation of original tissue dimensions without shrinkage is vital for stereology. Three steps, fixation, sampling and embedding, are essential requirements to minimise tissue shrinkage to obtain theoretically unbiased estimates of stereological parameters of lung structures. Perfusion fixation by intratracheal instillation with 1.5% glutaraldehyde/1.5% formaldehyde at a pressure of 25 cm fluid column is considered as one of the best methods. A systematic uniform random sampling scheme is then applied to the fixed lung to ensure each and every part of the lung is analysed, irrespective of homogeneity or heterogeneity of the structural distribution. The sampled tissue sections are then embedded in glycol methacrylate to minimise further tissue shrinkage. Here we describe the accurate fixation, sampling and embedding for stereological methods to quantify lung structures in mice.