Autonomous sailboats are promising platforms for long-term marine science missions and have become a research area of increased interest over the last two decades. To date, dozens of distinctive autonomous sailboats have been designed and notably employed in numerous tasks. Some literature lists and reviews the various designs of the existing autonomous sailboats; however, no comprehensive work connects the design to the performance requirements of various application scenarios. This paper first reviews and summarizes the existing designs from the perspective of critical performance in marine science missions, further pointing out the present state and the logic behind the designs. We then identify factors that hinder further performance improvement of autonomous sailboats through statistics and analysis of existing designs. We finally describe how the autonomous sailboat community should best address these challenges with technology from other disciplines. This article can provide references for designers of autonomous sailboats and inspire the community to eliminate the limitations they are facing. Additionally, making autonomous sailboats more powerful platforms can facilitate marine science research, such as research on ecosystems, biogeochemistry, and meteorology.