Purpose: The objective of this study is to develop a data middleware for u-IT convergence in agricultural environment monitoring, which can support non-standard data interfaces and solve the compatibility problems of heterogenous sensor networks. Methods: Six factors with three different interfaces were chosen as target data among the environmental monitoring factors for crop cultivation. PostgresSQL and PostGIS were used for database and the data middleware was implemented by Python programming language. Based on hierarchical model design and key-value type table design, the data middleware was developed. For evaluation, 2,000 records of each data access interface were prepared. Results: Their execution times of File I/O interface, SQL interface and HTTP interface were 0.00951 s/record, 0.01967 s/record and 0.0401 s/record respectively. And there was no data loss. Conclusions: The data middleware integrated three heterogenous sensor networks with different data access interfaces.