The purpose of this study was to examine empirically the relationship between productivity and size of coconut lands in Sri Lanka. The primary data collected by means of a field survey conducted by the Coconut Research Institute of Sri Lanka from 69 coconut cultivations located in the Kurunegala, Puttalam and Gampaha districts in 2004/05 were used. These cultivations were categorized into three groups depending on their size, namely: (1) Home gardens [less than 2 acres]; (2) Smallholdings [2 - 20 acres], and (3) Estate sector [more than 20 acres] for the purpose of analysis. First, the empirical model explaining the relationship between productivity and land size was tested with different functional forms (linear, quadratic, cubic, logarithmic), and next, several other variables that can have an impact on productivity (e.g. land suitability class, age of the plantation, number of bearing palms per acre, agro-ecological region, availability of labour and management practices adopted by growers such as fertilizer application, soil and moisture conservation and pest ands disease control etc.) were added into the selected model (i.e. cubic) and retested. The outcome of analysis shows that there exists a significant negative relationship between land size and productivity of coconut lands in Sri Lanka for home gardens and smallholders. The minimum productive land size was 21 acres, where the estimated productivity was 426 nuts / acre / year. On the other hand, the productivity of estates increases as the land size increases from 21 acre onwards up to 40 acres. Only the availability of labour was significantly affected on productivity, while the impact of other factors was insignificant in this respect. The results imply that policymakers and regulatory agencies responsible for coconut industry in Sri Lanka must take efforts to secure large estates from unnecessary fragmentation, and fragmentation of smallholding may be allowed if the separated lands are remained predominantly with coconut.