 3D printing technology has recently improved, but still faces limitations, and this paper proposes a new data-driven machine learning platform using multilayer perceptron and convolution neural network models to predict optimized parameters of the 3D printing process from model design to complete product. The proposed approach can quickly and accurately predict decisive parameters such as time, weight, and length while accounting for missing initial information and does not require knowledge of object shape, size, or material. A configurator is also proposed to simplify the 3D printing process by setting parameters for specific printer types. This article was authored by Fung Dong Ngyin, Tom Q. Ngyin, Q. P. Dao, and others.