Abstract: The experiment was conducted at the Agronomy Farm, Sylhet Agricultural University during November 2012 to February 2013 to know the performance of pea varieties under rice-based relay and conventional tillage cropping systems. There were two sowing methods; i. Relay with transplant aman rice and ii. Conventional through land preparation after harvest of transplant rice and four pea varieties; i. BARI motoshuti-1, ii. IPSA motorshuti-1, iii. Jhikorgacha local and iv. Natore local as treatments in the experiment. The experiment was conducted in a split plot design assigning sowing methods in the main plot and varieties in the sub-plots. The results revealed that the highest green pod yield (3333 kg/ha) was obtained from Natore local and this was statistically at par with that of Jhikorgacha local (3234 kg/ha) in the relay cropping system. The lowest green pod yield (1745 kg/ha) was obtained from BARI motorshuti-1 under conventional method of sowing. Due to lower number of pod/plant all varieties produced lower green pod yield in spite of higher plant population/m2 in the conventional sowing method. Natore local motorshuti produced the highest fodder yield (4055 kg/ha) under relay cropping system which was significantly different from the others and BARI motoroshuti-1 produced the lowest (2220 kg/ha) under conventional sowing method. The maximum gross return (Tk 1,74,760 /ha) was obtained from the variety Natore local followed by Jhikorgacha local (Tk 1, 69, 218 /ha) in relay cropping system. Similar trends were found in case of gross margin.