Nickel-based superalloy turbine blade castings are widely used as a key part in aero engines. However, due to the complex manufacturing processes, the complicated internal structure, and the interaction between different parts of the turbine blade, casting defects, such as stray grains, often happen during the directional solidification of turbine blade castings, which causes low production yield and high production cost. To improve the quality of the directionally solidified turbine blade castings, modeling and simulation technique has been employed to study the microstructure evolution as well as to optimize the casting process. In this article, a modified cellular automaton (MCA) method was used to simulate the directional solidification of turbine blade casting. The MCA method was coupled with macro heat transfer and micro grain growth kinetics to simulate the microstructure evolution during the directional solidification. In addition, a ray tracing method was proposed to calculate the heat transfer, especially the heat radiation of multiple blade castings in a Bridgman furnace. A competitive mechanism was incorporated into the grain growth model to describe the grain selection behavior phenomena of multiple columnar grains in the grain selector. With the proposed models, the microstructure evolution and related defects could be simulated, while the processing parameters optimized and the blade casting quality guaranteed as well. Several experiments were carried out to validate the proposed models, and good agreement between the simulated and experimental results was achieved.