We consider a search for the CP-violating angle delta(CP) in long baseline neutrino oscillation experiments. We show that the subleading delta(CP)-dependent terms in the nu(mu)->nu(e) oscillation probability can be easily obscured by the ambiguity of the leading term which depends on parallel to delta m(31)(2)parallel to. It is thus necessary to determine the value of delta m(31)(2) with a sufficient accuracy. The nu(mu) survival events, which can be accumulated simultaneously with the nu(e) appearance events, can serve for this purpose owing to its large statistics. Therefore, the combined analysis of nu(e) appearance and nu(mu) survival events is crucial to provide a restrictive constraint on delta(CP). Taking a test experimental setup, we demonstrate in the delta(CP)-delta m(31)(2) plane that the analysis of nu(e) appearance events leads to less restrictive constraints on the value of delta(CP) due to the ambiguity of delta m(31)(2) and that the combined analysis efficiently improves the constraints.