The phase behaviors and microstructures of various realistic and model mixtures of macro and micro molecules, such as polyolefin solutions and nanoparticle block copolymer composites, have been accurately predicted by the application of Statistical Associating Fluid Theory (SAFT) based approaches through various extensions that improve both the physical description of molecular interactions and efficiency of computations. The extensions are presented in a generic sense that is applicable to other studies. These rigorously derived theories have been demonstrated to capture material structure-property relationships and can be applied broadly to other fields including biology, medicine and energy industry. On the phenomenogical scale, the novel SAFT-Dimer equation of state has been extended to study the liquid-liquid phase boundary (cloud point) in polyolefin solutions. A simplified model of the polyolefin molecules has been followed and the effect of various parameters, such as temperature, molecular weight, solvent quality and comonomer content, on the phase behavior has been successfully captured by the theoretical model through comparison with experimental measurements. The presented approach requires less parameters than previous methods and is of critical value to the industrial productions of polymers, especially polyolefins with long branches. On the molecular scale, the interfacial SAFT (iSAFT) Density Functional Theory (DFT) has been extended to include a dispersion free energy functional that explicitly accounts for molecular correlations. The Order-Disorder Transition (ODT) between lamellar and disordered phase has then been investigated for pure block copolymer and copolymer nanocomposite systems. The extension has been shown to dramatically improve the ODT predictions of iSAFT as well as the self assembled microstructures in nanocomposites over previous DFT calculations, in comparison to coarse grained molecular simulations. The behavior of the equilibrium spacing of ordered structures is also examined against the variation of copolymer size and interactions. An efficient numerical scheme, Fast Fourier Transform (FFT), has been implemented and shown to drastically increase the computation efficiency. The theory has then been extended to study block copolymer morphologies with density variations in multiple dimensions. Comprehensive phase diagrams including lamellar, cylindrical and disordered phases have been obtained for copolymer nanocomposites for the first time using a single framework molecular theory. In addition, the nanoparticle induced morphological transition between cylindrical and lamellar phase has been studied using a pseudo arc-length continuation method. Transition evolution is tracked and metastable morphologies are examined and compared with existing experimental reports and theoretical calculations. With these extensions, iSAFT offers a powerful prediction tool that closely relates molecular structure to thermophysical properties and provides an efficient alternative to screen parameter space for specified material properties.