Flow visualization experiments were used to establish the mechanism of boiling on enhanced surfaces having sub-surface tunnels with surface pores. Based on this mechanism, a semi-analytical model for nucleate boiling is proposed. The model is validated for data covering a range of tunnel and pore dimensions and supporting bubble dynamics data. The model assumes the tunnel is vapor filled, except liquid menisci in the corners. By analyzing meniscus thickness, bubble departure diameter, and bubble growth, this dynamic model accounts for the temporal evaporation rate variation inside tunnels, and the dynamic nature of convection heat transfer outside the tunnels. The model uses two empirical constants. It predicts the heat transfer coefficient for the R-11, R-123, R-134a and R-22 pool boiling data with MSD = 20%. It also predicts bubble departure diameter and bubble frequency with MSD = 7 and 15%, respectively.