Illumina based 2nd generation sequencing provides us a powerful tool to understand transcriptome complexity. However, the limitation of short reads produced from next generation sequencing and reverse transcription application to covert RNA to cDNA for sequencing make the analyzed results not fully reliable because of false positive and false negative and lacking of RNA base modification information. Oxford nanopore sequencing technology can directly sequence full-length native RNAs which can accurately detect alternative splicing events and acquire RNA base modification information. With the technology, we produce more than 2 million reads of native poly(A)-tailed mRNAs from Arabidopsis seedlings and floral buds. Among these reads, around 36% of them are full-length transcripts judged by their full coverage of an existing transcript. Using these sequenced reads, we discovered more than 10000 novel isoforms not reported before, Moreover, we also found that m5C modification mainly distributed in the CDs region, and the m5C modification level between seedlings and floral buds is quite different, which may reflect different physiological status caused by different developmental stages.