This study presents a detailed investigation into conformational, physicochemical and adhesive characteristics of biopolymers on the surfaces of Escherichia coli and potential probiotic Bacillus subtilis harvested from middle-exponential phase (mid-EP), late-exponential phase (late-EP) and early-stationary phase (early-SP) of growth. The lengths to which bacterial surface biopolymers extend (biopolymer brush lengths), densities of grafted bacterial surface biopolymers indicating the amounts of molecules covering the bacterial surfaces (biopolymer grafting densities), adhesion forces of bacterial surface biopolymers to the model inert surfaces of silicon nitride (Si3N4), and the pull-off distances of biopolymers from Si3N4 were measured in water by atomic force microscopy (AFM). The Weibull analysis of AFM adhesion data showed that as the culture aged, the adhesive bonds between Si3N4 AFM tips and surface molecules of E. coli harvested from the culture were broken with a higher applied force. However, the highest applied force to break the bonds was required for B. subtilis in late-EP, followed by those required for cells in early-SP and mid-EP, respectively. The results of a steric model fitting to AFM approach force-distance (FD) curves and analysis of the pull-off distances in the AFM retraction FD curves showed higher biopolymer grafting density for E. coli in early-SP and longer biopolymer brush layer for B. subtilis in late-EP, which were associated with stronger adhesion to Si3N4 in water. The results of thermodynamic adhesion energy calculations based on the Wu model showed that polar interaction energy dominated the bacterial adhesion at the macroscale, the strength of which varied as a function of the growth phase for both E. coli and B. subtilis. The growth phase-dependent variation in polar components of thermodynamic adhesion energies between the bacteria and Si3N4 in water was consistent with the growth phase-dependent variation in the bond strengths between the bacteria and Si3N4 in water as revealed by the Weibull analysis of AFM adhesion data. Therefore, information obtained by Weibull analysis of nanoscale AFM bacterial adhesion data can be used to predict macroscale bacterial adhesion to the model inert Si3N4 surface in water. Display omitted •Higher biopolymer grafting density of E. coli in early-stationary phase led to stronger adhesion.•Longer biopolymer brush length of B. subtilis in late-exponential phase led to stronger adhesion.•Adhesion energy predictions of the Wu model agreed well with nanoscale adhesion measurements.•Weibull analysis of AFM adhesion data can be used to predict macroscale bacterial adhesion.