A bearing capacity problem for foundations on granular soil is often simplified by assuming a constant peak friction angle for the entire soil domain. This may lead to considerable deviation in estimating bearing capacity from actual field condition. Experiments for determining the shear strength of granular soil indicate that peak friction angle and dilation angle depend on the effective mean stress. Previous research works explored bearing capacity factor (Nγ) variation with foundation widths, accounting only the peak friction angle variation with stress, neglecting the dilation angle. Granular soil like sand is dilative. Therefore, the analysis considering only peak friction angle overestimates the bearing capacity of a foundation. Thus, it is crucial to include dilation angle effects in bearing capacity calculations. This study aims to establish a relationship between dilation angle, effective mean pressure, and Nγ variation for different foundation widths, using limit analysis and second-order conic programming to get the lower bound solution of the true collapse load. The study has been done by considering a surface strip footing resting on drained sand, subjected to vertical, inclined and eccentric loading. In the analysis, both loose and dense sands are considered as representative soils. Results of this analysis show that the Nγ value estimated by considering the dilatation effect becomes one-half to one-fourth of the Nγ value while not considering the effect of soil dilation. Hence, only accounting for peak friction angle variation with mean stress leads to strength overestimation, highlighting the necessity of incorporating dilation angle variations for accurate bearing capacity estimation.