Cell mechanics is an emerging field of research for translational medicine. Here, the cell is modeled as poroelastic cytoplasm wrapped by tensile membrane (poroelastic@membrane model) and is characterized by the atomic force microscopy (AFM). The parameters of cytoskeleton network modulus EC, cytoplasmic apparent viscosity ηC, and cytoplasmic diffusion coefficient DC are used to describe the mechanical behavior of cytoplasm, and membrane tension γ is used to evaluate the cell membrane. Poroelastic@membrane analysis of breast cells and urothelial cells show that non‐cancer cells and cancer cells have different distribution regions and distribution trends in the four‐dimensional space composed of EC, ηC. From non‐cancer to cancer cells, there is often a trend of γ, EC, ηC decreases and DC increases. Patients with urothelial carcinoma at different malignant stages can be distinguished at high sensitivity and specificity by analyzing the urothelial cells from tissue or urine. However, sampling directly from tumor tissues is an invasive method, may lead to undesirable consequences. Thus, AFM‐based poroelastic@membrane analysis of urothelial cells from urine may provide a non‐invasive and no‐bio‐label method to detecting urothelial carcinoma. This paper presents a method to characterize the mechanical properties of cells based on poroelastic@membrane model. The method can intuitively show the mechanical differences between non‐cancer cells and cancer cells and is expected to be applied to the diagnosis of urothelial carcinoma with a non‐invasive method in clinic.