The southeastern and southwestern parts of Bangladesh are naturally contaminated with arsenic exposing more than 30 million people to unsafe levels of this element in drinking water and potentially threatening rice production. A study was undertaken in three arsenic contaminated upazilas namely Kachua, Bhanga and Faridpur, aiming at understanding the possible negative effects of arsenic contamination on crop production. Sample survey was carried out to generate primary data. Two-stage sampling technique was followed in selecting the sample farms. Using Cobb-douglas yield function, productivity variability and the factors influencing level of productivity were identified. Analysis reveals that the share of agriculture income was higher for the farms under Bhanga (48%) compared to that under Kachua (46%). More than 70% of the sample households faced various arsenic related problems in rice production. Arsenic contamination resulted in less tillering, shorter plants, uneven plant growth and finally, decreased yield. Rice farmers adopted few practices for overcoming the problem, such as draining out of water from the rice fields, applying adequate fertilizers to the rice fields. Land degradation due to continued use of arsenic contaminated irrigation was reported. Due to the application of extra fertilizer and labour, the cost of modern Boro rice production in the more arsenic contaminated plots was 5% higher compared to that in less contaminated plots. Yields of MV (Modern variety) boro rice in more contaminated plots were significantly low resulting in lower gross return and profitability. Power tiller cost (a proxy for use of tiller), distance of plot from the STW, labour use, frequency of irrigation etc were the dominant determinants of MV Boro rice production in the sample arsenic prone areas.