Technological innovation and the "Belt and Road" initiative are closely linked and increasingly affect the high-quality and sustainable development of the modern economic system. However, the regions along the "Belt and Road" are still facing development difficulties such as uneven distribution of innovation resources, insufficient transformation of innovation results, imperfect construction of innovation ecosystems, and inadequate optimization of innovation systems and mechanisms. This restricts the improvement of regional scientific and technological innovation efficiency. To a certain extent, it affects the deepening development of the "Belt and Road" initiative. Facing the continuous deepening of the innovation-driven strategy and the continuous high-quality development of the "Belt and Road", clarifying the overall input and output efficiency of scientific and technological innovation in the provinces along the "Belt and Road", and exploring the factors that affect the efficiency of technological innovation in the provinces along the "Belt and Road" and how to improve the efficiency of scientific and technological innovation in provinces along the "Belt and Road", the analysis and resolution of these problems have attracted widespread attention from the practitioner and scholar. In this context, based on the three-stage data envelopment (DEA) model, according to panel data from 2012 to 2016, and taking full account of the time lag effect of the transformation of innovation achievements, we explore the real scientific and technological innovation efficiency of 17 provinces along the "Belt and Road" after excluding the external environmental impact. And the Tobit model is used to test the impact of technological innovation entities (high-tech enterprises, governments), labor quality, financial support and entrepreneurial level, and technological innovation policy factors on the efficiency of technological innovation in provinces along the "Belt and Road". The research results show that the efficiency of scientific and technological innovation in provinces along the "Belt and Road" is effective, but also inadequate, unbalanced and unstable. Firstly, the comprehensive efficiency of scientific and technological innovation in the provinces along the "Belt and Road" is spatially different. The comprehensive efficiency of scientific and technological innovation in the four major sectors of the Southeast, Northeast, Southwest and Northwest has obvious differences, and the overall efficiency is higher in the east and lower in the west. Secondly, the comprehensive efficiency of scientific and technological innovation in provinces along the "Belt and Road" has dynamic changes over time. The overall efficiency has shown an upward trend in the past five years, and the gap in comprehensive efficiency between the four major sectors has decreased year by year. Finally, the analysis of the impact of the Tobit regression model on the overall efficiency of scientific and technological innovation shows that government support, labor quality, and financial environment have a positive impact on the efficiency of scientific and technological innovation. The intensity of scientific research expenditures, the "Belt and Road" initiative policy and the level of entrepreneurship have a negative impact on the efficiency of scientific and technological innovation. Based on this, we believe that it is reasonable to increase the efficiency of scientific and technological innovation in provinces along the "Belt and Road" in terms of rationally deploying scientific and technological innovation resources, cultivating high-quality innovation ecosystems, optimizing innovation systems and mechanisms, and improving scientific and technological innovation policy systems.