An H−/H∞ sensor fault detection and isolation problem is investigated for continuous- and discrete-time Takagi–Sugeno fuzzy systems with state delay and immeasurable premise variables. A bank consisting of the sensor’s number of fuzzy observers is adopted. A fuzzy observer gain and a fuzzy residual gain in each observer are designed such that the residual is sensitive to a certain partial group of faults and robust against disturbance. Sufficient design conditions are derived in nonlinear matrix inequality format and a numerically tractable algorithm involving a convex optimization is presented based on the cone complementary linearization technique. A simulation is provided to verify the effectiveness of the proposed technique.