Melanoma is an aggressive type of skin cancer. Current methods for diagnosis rely on visual inspections, which are highly subjective and depend on the physician’s ability. The objective of this study was to detect key differences in the dielectric permittivity between normal skin cells and melanoma using a tissue phantom with an electromagnetic skin patch sensor. An oil-in-gelatin skin phantom was fabricated to mimic dielectric properties of skin tissue, melanoma, and subcutaneous fat. An electromagnetic resonant sensor was fabricated as a square planar spiral and placed on the tissue phantom to collect data using a vector network analyzer on both normal cells and melanoma. Frequency sweep data including the S11 reflection coefficient was generated for each measurement. The mean first principal resonant frequency was measured at 757MHz for normal skin tissue and at 710MHz for melanoma, thus indicating a frequency downshift of 47MHz. An ANOVA was performed (p<0.0001) indicating significant changes between normal and melanoma tissue. These changes in resonant frequency suggest that the difference in dielectric properties between normal cells and melanoma could be used to screen melanoma with a novel biosensor. The implementation of a low-cost diagnostic skin patch in melanoma populations may produce a strong costs effective solution for the melanoma healthcare expenditure of a nation.