In this paper, an easily implementable immersed interface method based high-resolution phase-field model is proposed for simulating dendritic growth related problems in complicated geometries. At first, resolution issues of different Fourier Pseudo-spectral based dealiasing schemes are shown for the anisotropic phase field equations at high thermal undercooling conditions using just adequate spatial grid resolution. The problems of aliasing error and spectral leakage are studied using high-order based exponential smoothing spectral filters, which generally occur at low sampling rate situations. A fully dealiased zero padding scheme based anisotropic phase-field method is also developed in conjunction with an optimal third order strong stability preserving Runge–Kutta (SSPRK3) time integration scheme for obtaining more accurate and stable numerical solutions. In the later part, a modified phase-field method is implemented with the indicator function which simulates different crystal growth problems in complex geometries without using non-trivial mesh refinement algorithm. Effect of non-interacting obstacles present in the computational domain can also be readily incorporated in the present numerical model by using the single order parameter based phase-field equations.