Technological innovation projects play a crucial role in driving economic growth and societal progress. Evaluating the feasibility and potential of such projects is essential for decision-makers in various industries. With the advent of big data analysis and machine learning, there is an opportunity to develop advanced algorithms that can leverage the vast amount of information available to make informed assessments. This paper focuses on the application of big data analysis and machine learning algorithms, particularly neural networks, as a means of evaluating technological innovation projects. These algorithms are capable of processing and analyzing large datasets, extracting valuable insights, and predicting project outcomes. In order to provide an assessment model for technical innovation initiatives, this study is built on BP. First, considering that fireworks algorithm (FWA) is prone to falling into local optimum, this study suggests improvement approaches to build IFWA. Then, in order to build IFWA-BP, this study employs IFWA to optimize the basic network parameters. In order to improve performance, this technique can address the drawbacks of typical BP networks. Lastly, the suggested approach is tested systematically, and the findings show that the procedure is reliable.