Motivation: Improving the prediction of structures, especially those containing pseudoknots (structures with crossing base pairs) is an ongoing challenge. Current alignment-based prediction algorithms only find the consensus structure, and their alignments can come from structure-based alignment algorithms, which is more reliable, but come with an increased cost compared to sequence-based alignment algorithms. This step can be removed; however, non-alignment based algorithms neglect structural information that can be found within similar sequences. Results: We present a new method for prediction of RNA pseudoknotted secondary structures that combines the strengths of MFE prediction and alignment-based methods. KnotAli takes an RNA sequence alignment and uses covariation and thermodynamic energy minimization to predict secondary structures for each individual sequence in the alignment. We compared KnotAli's performance to that of three other alignment-based algorithms, on a large data set of 10 families with pseudoknotted and pseudoknot-free reference structures. We produced sequence alignments for each family using two well-known sequence aligners (MUSCLE and MAFFT). We found KnotAli to be superior in 6 of the 10 families for MUSCLE and 7 of the 10 for MAFFT. We find KnotAli's predictions to be less dependent on alignment quality. In particular, KnotAli is shown to have more accurate predictions compared to other leading methods as alignment quality deteriorates. Availability: The algorithm can be found online on Github at https://github.com/mateog4712/KnotAli / Graduate / 2022-08-16