Here we develop BioNavi-NP, a navigable and user-friendly toolkit which are capable of predicting the biosynthetic pathways for NP and NP-like compounds, using a novel learning (AND-OR Tree)-based planning algorithm and enhanced molecular Transformer neural networks. For each biosynthetic step in the multi-step bio-retrosynthetic routes, we further evaluate the plausible enzymes through enzyme prediction tools. BioNavi-NP is promising for rationally reconstructing the complex biosynthetic pathways and efficiently exploring the reaction space of NPs. The package can be found at Github and please make use of the downloadable version when the server is busy. (Zheng et al. Nat. Commun. 2022)