We discuss primary CR energy estimations for very large surface EAS experiments. We show that the procedure of energy determination via the relation to particle densities at 600 m from the shower axis, ρ, requires a specific treatment for inclined cascades when the maximum of EAS development is close to the array. As follows from simulations and cascade theory, using an exponential function for ρ conversion from inclined to vertical showers is no longer valid and leads to an overestimate of the primary energy. We present a procedure to reconstruct the primary spectrum that can reduce the discrepancy between the results of fluorescence experiments and surface arrays.