Smart meter data analysis provides key insightsabout energy demand and usage patterns for efficient operationof power generation and distribution companies. The increasein modern communication bandwidth enables smart meters totransmit the data to a corresponding utility company at hourlyupdate rates or faster. Analysing such large amount of data often requires a highperformance cloud computing environment. However, usingsuch environment may lead to exposure of energy consumptionpatterns of individual households, with the potential consequenceof damaging privacy breaches. To mitigate the risk of a privacy breach, this paper proposesa secure linear regression model for smart meter data analytics, based on a Partially Homomorphic Encryption algorithm. Inthe proposed method, the primary variable, here, the powerreading, is encrypted. The statistical coefficients are thencomputed directly from the cyphertext using integer mappings. With this approach, a computationally feasible linear regressionis achievable without compromising a detailed householdenergy usage profile. Simulation experiments are conductedthat demonstrate the performance of proposed method withrespect to accuracy and computational complexity.